Client Network Traffic With Exchange 2000

Client Network Traffic with Exchange 2000
Client Network Traffic

with Exchange 2000
Authors :
Christophe LEROUX
Christophe BESANÇON
Microsoft Consulting Services – France
Version 1.0
August 8, 2000
The example companies, organizations, products, people, and events depicted

herein are fictitious. No association with any real company, organization, product,
person or event is intended or should be inferred.
2 Client
Network Traffic with Exchange 2000
Contents
1. Acknowledgments 6
2. Document Vision / Scope 7
3. Test Lab Configuration 8
3.1. Lab design 8
3.2. LAN design 9
3.3. Servers Characteristics 10
3.4. Client Characteristics 11
3.5. Exchange 2000 Configuration 11
3.6. Terminal Services Configuration 13
3.7. Client Configuration 13
3.7.1. Outlook 2000 Client Configuration 14
3.7.2. Outlook 97 Client Configuration 14
3.7.3. Outlook Express Client Configuration 14
3.7.3.1. POP3 Account Settings 15
3.7.3.2. IMAP4 Account Settings 15
3.7.3.3. Lightweight Directory Access Protocol
(LDAP) Account Settings 15
3.7.3.4. Network News Transfer Protocol (NNTP)
Account Settings 15
3.7.4. Netscape Messenger Client Configuration 16
3.7.4.1. POP3 account settings 16
3.7.4.2. IMAP4 account settings 16
3.7.4.3. NNTP account settings 16
3.7.5. Outlook Web Access Client Configuration 17
3.7.6. Terminal Services Client Configuration 17
3.8. Measurement Methodology 18
4. Log On and Log Off 20
4.1. Tests Performed 20
4.2. MAPI Clients: Microsoft Outlook 2000, Outlook 97 20
4.2.1. Test Details 21
4.2.2. Outlook 2000 results 24
4.2.4. Outlook 2000 and Outlook 97 measurements
Analysis 25
4.3. Microsoft Outlook Express: POP3 and IMAP4 Modes 26
4.3.1. Test details 26
4.3.2. Outlook Express IMAP4 Results 27
4.3.3. Outlook Express POP3 results 28
Client Network Traffic with Exchange 2000 3
4.3.4. Outlook Express Measurements Analysis 28

4.4. Netscape Messenger: POP3 & IMAP4 Modes 29
4.4.2. Netscape Messenger IMAP4 Results 30
4.4.3. Netscape Messenger IMAP4 measurements
Analysis 30
4.5. Outlook Web Access 30
4.5.2. Outlook Web Access Results 31
4.5.3. Outlook Web Access Measurements Analysis 31
5. Directory Access 32
5.2. MAPI Clients: Outlook 2000, Outlook 97 33
5.2.4. Outlook 2000 and Outlook 97 Measurements
Analysis 40
5.3. Outlook Express: LDAP mode 42
5.3.2. Outlook Express LDAP Results 43
5.4. Netscape Messenger: LDAP mode 43
5.5.3. Outlook Web Access Measurements Analysis 45
6. Generic Mail Item Tests 46
6.1. Tests performed 46
6.2. Test Details 47
6.3. Outlook 2000 Results 50
6.5. Outlook Express IMAP Results 57
6.6. Outlook Express POP Results 62
6.7. Netscape Messenger IMAP Results 66
6.8. Netscape Messenger POP Results 71
6.9.1. Tests Details 75
4 Client
6.10. Mail Item Analysis 80

7. Calendaring, Contacts and Task items Tests 87
7.5. OWA Results 91
7.6. Calendaring, Contacts, and Tasks Analysis 92
8. Public Folders 94
8.3. Outlook 2000 results 96
8.4. Outlook 97 results 102
8.5. Outlook Express NNTP Results 104
8.6. Netscape Messenger NNTP Results 109
8.7. OWA results 114
8.8. Global results 115
9. Outlook 2000 with Terminal Services 120
9.1. Introduction 120
9.3. Terminal Services Client Results 126
9.4. Terminal Services Measurement Analysis 128
10. Web Storage System 129
10.3. Log On and Log Off and User Traffic Results 130
10.4. User traffic 131
10.5. Web Storage System Measurement Analysis 132
11. Instant Messaging 133
11.3. Log On and Log Off 134
11.4. User Traffic Results 134
11.5. Global results 136
11.5.1. Status Change 136
11.5.2. Sending Messages 136
12. Client Traffic Measurement Conclusions 137
13. How client use Front-End / Back-End topology ? 138
13.1. Test Lab Configuration 138
13.1.1. Lab design 138
13.1.2. LAN design 139
13.1.3. Servers characteristics 140
13.1.4. Clients characteristics 141
13.1.5. Exchange 2000 configuration 141

14. Front-End / Back-End Communications 146
14.1. Message Sending 146
14.1.1. Analysis 150
14.2. Message reading 151
14.3. Global analysis about Front-End / Back-End
communications 158
15. Routing Group Connector 159
15.1. Lab configuration 159
15.2. Message Sending 161
16. Appendix 170
16.1. DSProxy 170
16.2. Client Access to Active Directory 171
16.2.1. Outlook 98 and Outlook 2000 171
16.2.1.1. First discovery of the GC the Outlook
2000 client 173
16.2.1.2. Other sessions opened with the same user
profile with Outlook 2000 183
16.2.2. Exchange Client and Outlook 97 and
Outlook 98 183
16.2.3. Traffic and Load Generated Through the
DSProxy Process 183
17. Index 185
17.1. Graphs 185
17.2. Figures 188
6 Client
1. Acknowledgments
Many thanks to the following people for their time and enthusiasm. No white paper
is produced without many people going out of their way to assist.
Microsoft Consulting Services, France

Frédéric Arnoux
Hervé Captier
Olivier Douetteau
Marc Duvert
Marie-Anne Loulié-Tuquet
Philippe Martineau
Patrice Trousset
François Valraud
FR-PSS REC Biz Apps

Damien Caro (SRG ENGINEER)
Deployment Assets Team

Matthias Leibmann
Andreas Luther
Other
Marie-Laure Leroux-Angibaud
2. Document Vision / Scope
Deploying Microsoft Exchange 2000 Server in a company requires a thorough

understanding of a company’s structure, network topologies, usage patterns, and so
on. A particularly difficult task is estimating the traffic generated between multiple
Microsoft Exchange 2000 servers and messaging clients.
This section provides a detailed look at how Microsoft Exchange 2000 consumes
bandwidth under different configurations and analyzes the traffic between clients
and Microsoft Exchange 2000 Server.
System professionals can use the statistics and information to better understand
network bandwidth loading, including how to set up segments, assign users, and
extrapolate the results for various configurations.
The traffic analysis in this section is based on the results of 700 tests conducted
with the following messaging clients:
• Microsoft Outlook 2000
• Microsoft Outlook Express 5.0
• Microsoft Outlook Web Access
• Microsoft Exchange Instant Messaging 2.0
• Netscape Communicator 4.7
Test data appears in Microsoft Excel workbooks and on the Exchange 2000 Server
Resource Kit companion compact disc (CD). In some cases, the data from these
tests do not represent all information required by the clients. Network bandwidth is
not consumed for graphics caching, session lifetimes, resolved names caches, and
message formats. However, message content and form significantly impacts
network traffic, especially an HTML message. These tests allow you to compare
clients and mail message format Rich Text Format (RTF), HTML, Plain Text. You
can also use the data to estimate the network traffic generated by many user
profiles when their messaging habits are well known.
Microsoft Outlook 98 was not tested because its performance is similar to the
performance of Outlook 2000. For more information about testing Outlook 98, see
"Directory Access" in this section.
8 Client
3. Test Lab Configuration
3.1. Lab design
The test lab was built with four computers running Windows 2000 Server. One
server ran Exchange 2000 Server RC2 (build 4368) and the others ran
Windows 2000 services like domain controller, global catalog, and Domain Name
System (DNS).
The three client computers had the following products installed:
• Outlook 2000 (version 9.0.0.2711)
• Outlook 97 (version 8.04.5619)
• Outlook Web Access (with Microsoft Internet Explorer 5
• Outlook Express 5.0 (version 5.00.2919.67.00)
• Netscape Messenger 4.7
• Exchange Instant Messaging 2.0 (version 2.0.1002)
• Terminal Services client (32-bit version)
Figure 1 : Network Topology
WINDOWS 2000 Server

Domain Controller
Global Catalog Server
London 01 WINDOWS 2000 Professional
Outlook 2000
Outlook Express 5.0
Instant Messaging client
London 05
Internet Explorer 5.0
WINDOWS 2000 Server

Domain Controller
London 02
WINDOWS 2000 Professional

Netscape Messenger 4.7
WINDOWS 2000 Server Lonbdon 06

Member Server
DNS Service
Network Monitor Lodon 03
Outlook 97
Outlook Express 5.0
Instant Messaging client
WINDOWS 2000 Server London 07 Terminal Services client
Member Server
Exchange 2000 Server
Terminal Services
London 04
Outlook 2000
3.2. LAN design
The DNS root domain is called microsoft.com. The server, LONDON-03, manages
the DNS primary zone and allows dynamic updates. Every server has a static
Internet Protocol (IP) address and refers to the DNS to resolve names.
To avoid disturbing the tests and experiencing caching phenomena, the media
access control (MAC) and IP addresses are preloaded on the client computers. This
prevents Address Resolution Protocol (ARP) or DNS queries from being included
in the test results.
The network is an Ethernet 100–megabyte (MB) dedicated network.
10 Client
Table 1 : Server configurations
Server IP address IP Subnet mask Role

LONDON-01 10.0.0.1 255.0.0.0 Domain controller
Global catalog
LONDON-02 10.0.0.2 255.0.0.0 Domain controller
LONDON-03 10.0.0.3 255.0.0.0 Member server
DNS
LONDON-05 10.0.0.5 255.0.0.0 Member workstation
3.3. Servers Characteristics
The following table details the processor speed, memory, and software that runs on
each server.
Table 2 : Server characteristics
Server CPU Memory OS Other

Software
LONDON-01 Intel Pentium II 128 MB Windows 2000
450 MHz Advanced Server
LONDON-02 Intel 486 66 MHz 64 MB Windows 2000
Advanced Server
LONDON-03 Intel Pentium III 128 MB Windows 2000
500 MHz Advanced Server
LONDON-04 Intel Pentium III 128 MB Windows 2000 Exchange 2000
500 MHz Advanced Server Server
LONDON-08 Intel Pentium III 128 MB Windows 2000 Terminal
500 MHz Advanced Server Services
3.4. Client Characteristics
each client computer.
Table 3: Client characteristics
Client CPU Memory Operating Other

System Software
LONDON-05 Intel Pentium 166 144 MB Windows 2000 Outlook 2000
MHz Professional
Outlook
Express 5.0
Instant
Messaging 2.0
LONDON-06 Intel Pentium 166 144 MB Windows 2000 Netscape
MHz Professional Messenger 4.7
LONDON-07 Intel Pentium III 128 MB Windows 2000 Outlook 97
500 MHz Professional
Outlook
Express 5
Instant
Messaging 2
Terminal
Services client
3.5. Exchange 2000 Configuration
These tests capture network traffic between a client and a server. Use only one
Exchange 2000 server for this set of tests.
Install the server in the default routing group as a custom setup, because Instant
Messaging Service is also installed.
The following screen shot summarizes the installed components on this server.
12 Client
Figure 2 : Exchange 2000 installed components (1/2)
Figure 3 : Exchange 2000 installed components (2/2)

The messaging organization is Litware, Inc. (a fictional company), and the server
is in the default administration group, First Administrative Group. The mailbox
and public folder stores appear in the only storage group, First Storage Group. No
server options, mailbox store, or public folder store options have been changed; all
of the settings are defaults.
Create users in the default Windows 2000 organization unit as Users. The test
users have a mailbox, an e-mail address, and no password. Because Exchange
Instant Messaging demands a password, only Instant Messaging users have a
password.
3.6. Terminal Services Configuration
The LONDON-08 server is running Terminal Services and Terminal Services

Licensing. This enables the multi-user environment to access the server; thus, a
user can open a session remotely and run any installed server application. The
server evaluates the network traffic when using Outlook 2000 in the Terminal
server environment.
Figure 4 : Terminal service configuration
3.7. Client Configuration
For the tests, use the default client configuration defined by Setup. However, you
can alter some settings to reduce disruption in captures or to make the client
comparisons easier.
14 Client
3.7.1. Outlook 2000 Client Configuration

The Outlook 2000 default MAPI test profile included only two services:
Microsoft Exchange 2000 Server and Microsoft Outlook Address Book.
The following client settings were cleared before testing:
• Empty the Deleted Items folder upon exiting
• Save copies of messages in Sent Items folder
• AutoSave
• Preview Pane
The folder list is enabled only for the tests on public folders.
3.7.2. Outlook 97 Client Configuration

The MAPI test profile included only three services: Microsoft
Exchange Server, Outlook Address Book and Personal Address Book.
The following client settings were cleared:
• Empty the Deleted Items folder upon exiting
• Save copies of messages in Sent Items folder
The folder list is enabled for the tests on public folders.
3.7.3. Outlook Express Client Configuration

The following Outlook Express settings were cleared:
• Check for new messages every x minutes
• Automatically log on to MSN Messenger Service
• Save copy of sent messages in Sent Items folder
• Purge deleted messages when leaving IMAP folders
The settings Send and receive messages at startup and When starting, go
directly to My Inbox folder will be described in each test.
The HTML format is embedded in Multipurpose Internet Mail Extensions
(MIME) format and encoded with Quoted Printable.
The Plain Text format is encoded as MIME with Quoted Printable.
The Post Office Protocol version 3 (POP3) and Internet Message Access
Protocol version 4 (IMAP4) accounts are created alternatively, so that the
client never supports more than one mailbox.
3.7.3.1. POP3 Account Settings

• The setting Include this account when receiving mail or synchronizing is
selected
• The POP3 and Simple Mail Transfer Protocol (SMTP) servers are LONDON-
04.
• The setting Log on using Secure Password Authentication is cleared.
• The setting Leave a copy of message on server is cleared.
3.7.3.2. IMAP4 Account Settings

• The setting Include this account when receiving mail or synchronizing is
selected.
• The IMAP4 and SMTP server is LONDON-04.
• The setting Log on using Secure Password Authentication is cleared.
• The setting Break apart messages larger than is cleared.
3.7.3.3. Lightweight Directory Access Protocol (LDAP) Account

Settings
• The LDAP server is LONDON-01.
• The settign This server requires me to log on is selected.
• The account name is cn=alias,cn=microsoft.
(“cn=alias,cn=microsoft.com”, “cn=alias,dc=microsoft.com” also works).
• The setting Log on using Secure Password Avuthentication is selected;
otherwise, the log on will fail.
• The box Search base contains: dc=Microsoft, dc=com.
• The setting Use simple search filter is cleared.
3.7.3.4. Network News Transfer Protocol (NNTP) Account

Settings
• The NNTP server is LONDON-04.
• The setting This server requires me to log on is cleared. The default virtual-
NNTP server allows anonymous access.
Note Create the NNTP account only with the POP3 account.
IMAP4 allows native connection to public folders; POP3 does not
support native connection.
16 Client
3.7.4. Netscape Messenger Client Configuration

The following options under Mail and Newsgroups are disabled:
• Addressing, Address Books, Directory Server.
• All options in Copies and folders .
• All options in Formatting .
The following options are enabled:

• In Messages , Send messages that use 8-bit characters, using the Quoted
Printable MIME encoding.
• In Formatting, Send the message in HTML anyway.
Create the POP3 and IMAP4 accounts alternatively so the client never
supports more than one mailbox. Netscape Messenger does not allow a POP3
account while there is an IMAP4 account; it does, however, allow multiple
IMAP4 accounts.
3.7.4.1. POP3 account settings

• The POP3 and SMTP server is LONDON-04.
• The option Check for mail every x minutes is disabled.
• The option Leave a copy of message on server is enabled by default.
3.7.4.2. IMAP4 account settings

• The IMAP4 and SMTP server is LONDON-04.
• The option Check for mail every x minutes is disabled.
• The option Clean up Inbox on exit is disabled.
3.7.4.3. NNTP account settings

• The NNTP server is LONDON-04.
• The option Only ask me for my user name and password when necessary
is enabled. The logon is anonymous.
Note Netscape Messenger creates an NNTP account automatically.
Because accessing public folders is possible with IMAP4, use the
NNTP account with the POP3 account.
3.7.5. Outlook Web Access Client Configuration

In Outlook Web Access tests, we chose to use Microsoft Internet Explorer 5 to
get the benefits of the XML support.
The amount of network traffic varies, depending on whether or not graphics
are enabled on the browser. When graphics are enabled and present on HTML
pages, Internet Explorer searches for graphics in its cache folder before
downloading. If the action has already been done, the graphic is cached
therefore the traffic is lesser. We take numbers to reflect a normal use of OWA
client, where we perform many times the same action. So, these numbers do
not include traffic due to graphics except specific tests.
3.7.6. Terminal Services Client Configuration

The Terminal Services 32-bit client is installed on a workstation running
Windows 2000 Professional.
Figure 5 : Terminal Services Client
The default Terminal Services client parameters are illustrated in Figure 5. The
screen area is set to 800 x 600 pixels and the Enable data compression check
box is selected.
18 Client
3.8. Measurement Methodology
The network traffic captures were done with Microsoft Network Monitor 2.0
version 5.00.943, found in Microsoft System Management Server 2.0 Service
Pack 2.
Figure 6 : Network Monitor Capture Filter
A static network address resolution was established prevent unexpected network

broadcasts for MAC addresses or requests from DNS to get IP addresses from
name, both of which cause timeouts.
A batch file containing the command Arp.exe –s IP address Mac address was
created for each computer in the lab network then was run once on each computer.
It was build as below:
ARP.EXE –s 10.0.0.1 01-02-03-qb-cd-ef-99
A host file that contained all computer names and their IP address was also created
on each computer.
Each data capture was performed according to the following steps:
1. Start capture.
2. Perform the action to measure.
3. Wait until there is no more traffic.
4. Stop capture.
5. Save the capture.
6. Report results.
Each capture was performed two or three times for results comparison and
consistency checks.
The test messages had the following characteristics.
• For all the tests, except the public folders tests:
• Messages of 1 kilobyte (KB), 2 KB, and 4 KB were created, their sizes
determined by the size appearing on Outlook 2000 when the messages were
sent using RTF. These messages were saved as .msg files and their size on
the hard drive was greater than when they were stored in the mailbox. On
the hard drive, sizes were 5 KB, 10 KB, and 20 KB, respectively. These
messages contained a few characters per row and many carriage returns
(message content impacts network traffic when the messages are converted
into HTML format). The font in each message was Arial. The font style was
regular, and the font size was 10 points.
• Messages with attachments: had Microsoft Word or Microsoft PowerPoint files

of 10 KB, 50 KB, 100 KB, 500 KB, and 1,000 KB attached to the 1–KB
message described earlier.
• Public folders tests and NNTP access:

• Messages for public folders tests contained 1024, 2048, or 4048 characters,
creating messages of 1 KB, 2 KB, and 4 KB respectively. These non-RTF
messages were saved as .txt files on a hard disk drive. The text contained a
portion of the Microsoft Excel 97 Readme.txt and it contained no more than
75 characters per line.
• Messages with attachments: Microsoft Word or Microsoft PowerPoint files of

10 KB, 50 KB, 100 KB, 500 KB, and 1,000 KB, with the text of the 1–KB
message (1024 characters) described earlier.
20 Client
4. Log On and Log Off
These tests measure the network traffic generated by initial client connection and
validation on the Microsoft Exchange 2000 Server when users access their
mailboxes. Because each client is unique, the tests are partitioned to the following
four groups.
• MAPI clients: Microsoft Outlook 2000, Outlook 97
• Microsoft Outlook Express: POP3 and IMAP4 modes
• Netscape Messenger: POP3 and IMAP4 modes
• Web client: Microsoft Outlook Web Access
4.1. Tests Performed
Log On
This test captured the traffic generated by initial client connection and validation
and that of Microsoft Exchange Server to the global catalog server and domain
controller. For MAPI clients (Outlook 2000 and Outlook 97), the traffic was
captured from when Outlook was launched until the traffic dissipated.
Log Off
This test captured the traffic resulting from a disconnection from the server.
4.2. MAPI Clients: Microsoft Outlook 2000,

Outlook 97
For MAPI clients, the generated traffic depends on various parameters:

• Whether Outlook is running for the first time on the computer. If so, one or
more welcome messages arrive in the mailbox.
• Whether the mailbox is new. Mailboxes created on a server are not formatted
for Outlook. When you connect with Outlook to a new mailbox, extra traffic is
generated to initialize the mailbox on the server-that is, to create system folders
(Calendar, Contacts, Drafts, Journal, Notes, Tasks) and system views. Only
Outlook 2000 creates a Drafts folder.
• Whether the MAPI profile is new. When a MAPI profile is new, extra traffic is
generated to initialize it with settings stored in the registry for domain names of
the resolved name, the mailbox servers, and public folder servers.
• Whether the preview pane view is on. Only Outlook 2000 has this feature. The
preview pane is activated by default. Although the preview pane view can be
added to Outlook 97, the tests described in this appendix did not include
Outlook 97 with the preview pane view.
• Whether default folder contains items. The item count in the default folder has
an impact on traffic. Captures were done with inboxes that contained 50 and then
100 unread items.
4.2.1. Test Details

The following tests were performed.
LOGON1: First startup, new mailbox, new profile
1. Create a new mailbox on the Exchange server.
2. Install Outlook on the client computer.
3. Create a profile with the server name and user name.
4. Start Outlook.
To capture the traffic, the user name was not resolved.
LOGON2: First startup, existing mailbox, new profile

1. Install Outlook on the client computer.
3. Start Outlook.
In order to capture the traffic, the name was not resolved.

A MAPI client used the mailbox, but it was empty.
LOGON3: New mailbox, new profile

1. Create a new mailbox on the Exchange server.
3. Start Outlook.
To capture the traffic, the name was not resolved.

22 Client
LOGON4: Existing mailbox, new profile

In this test, a MAPI client had used the mailbox, but the mailbox was empty.
2. Start Outlook.
LOGON5: Existing mailbox, existing profile, no message

This is a common situation. A profile and mailbox exist, and a MAPI client has
used the mailbox already, but the mailbox is now empty. In this test, the preview
pane was disabled.
1. Start Outlook.
LOGON6: Existing mailbox, existing profile, 50 messages

In this test, the preview pane was disabled so that the size of the messages did not
affect the capture, and so that results between Outlook 2000 and Outlook 97 could
be compared.
1. Start Outlook.
2. Send 50 messages with 1024 characters as content.
3. Exit and log off Outlook.
4. Start Outlook.
Inbox contains 50 unread messages.
LOGON7: Existing mailbox, existing profile, 100 messages

In this test, the preview pane was disabled so that the size of the messages did not
affect the capture, and so that results between Outlook 2000 and Outlook 97 could
be compared.
1. Start Outlook.
2. Send 100 messages with 1024 characters as content.
3. Exit and log off Outlook.
4. Start Outlook.
Inbox contains 100 unread messages.
LOGON8: First startup, existing mailbox, new profile, no preview pane

This test was done with Outlook 2000 to measure the preview pane’s affect on the
network. This test was done on an existing mailbox.
1. Run Outlook 2000.
2. In Inbox, disable preview pane.
3. Uninstall Outlook 2000.
4. Install Outlook 2000.

6. Run Outlook
LOGOFF: Logoff
1. Close all Outlook windows except the main one.
2. Close Outlook by clicking File, and then click Exit and Log Off.
24 Client Network Traffic with Exchange 2000
4.2.2. Outlook 2000 results
Client to Exchange Exchange 2000

Total Sent by Client Received by Client Client to GC GC to Client
2000 Server Server to Client
Bytes Frames Bytes Frames Bytes Frames Bytes Frames Bytes Frames Bytes Frames Bytes Frames
LOGON1 164784 550 85546 289 79238 261 1904 19 83642 270 1472 14 77766 270
LOGON2 161218 529 83764 280 77454 249 1904 19 81860 261 1472 14 75982 261
LOGON3 56852 329 27068 177 29784 152 1904 19 25164 158 1472 14 28312 158
LOGON4 53576 311 25404 169 28172 142 1904 19 23500 150 1472 14 26700 150
LOGON5 24618 163 12246 93 12372 70 4034 40 8212 53 5036 30 7336 40
LOGON6 26314 163 12182 93 14132 70 4034 40 8148 53 5036 30 9096 40
LOGON7 27136 164 12182 93 14954 71 4034 40 8148 53 5036 30 9818 41
LOGON8 146482 522 73876 228 72606 294 1904 19 77608 209 1472 14 71134 180
LOGOFF 3460 33 1804 17 1656 16 508 6 1296 11 508 6 1148 10
Total Sent by Client Received by Client
Bytes Frames Bytes Frames Bytes Frames

LOGON1 50130 281 22708 149 27422 132
LOGON2 47838 272 21534 145 26304 127
LOGON3 50276 283 22784 150 27492 133
LOGON4 47248 268 21180 143 26068 125
LOGON5 23546 152 11450 84 12096 68
LOGON6 25830 153 11510 85 14320 68
LOGON7 27964 154 11510 85 16454 69
LOGOFF 3152 31 1654 16 1498 15
4.2.4. Outlook 2000 and Outlook 97 measurements

Analysis
Graph 1 : Logon (comparative: Outlook 2000, Outlook 97)
180 000
160 000
140 000
120 000
100 000
Bytes
80 000
60 000
40 000
20 000
0 Logon 1 Logon 2 Logon 3 Logon 4 Logon 5 Logon 6 Logon 7
Outlook 2000 164 784 161 218 56 852 53 576 24 618 26 314 27 136
Outlook 97 50 130 47 838 50 276 47 248 23 546 25 830 27 964
Logon1 = first run w/ welcome message, new mailbox, new profile.

Logon2 = first run w/ welcome message, new profile.
Logon3 = new mailbox, new profile.
Logon4 = existing mailbox, new profile.
Logon5 = existing profile.
Logon6 = 5 messages in Inbox.
Logon7 = 10 messages in Inbox.
26 Client
The measurements captured when running Outlook 2000 and Outlook 97 for the
first time are not comparable. The number and size of welcome messages and the
preview pane (only present in Outlook 2000) explains the difference.
The creation of a new profiles or new mailboxes generates more traffic. A new
profile generates 30 KB more traffic. The initialization of a new mailbox generates
3 KB more traffic. In daily use (where there is an existing mailbox and profile),
Outlook 2000 generates as much traffic (in bytes) as Outlook 97.
As more fields appear, such as previews of unread messages, network traffic
increases. The size of the message does not affect the traffic.
A close analysis of the data shows that Outlook 2000 contacts the closest global
catalog server, whereas Outlook 97 does not. Outlook 97 contacts the Exchange
server, which then sends the directory look-up requests to the global catalog server.
The global catalog server is contacted only during the logon process and during
address book-related queries.
Both versions of the Outlook clients generate equal network traffic when closing.
Outlook 2000 and Outlook 97 consume the same bandwidth during logon and
logoff.
4.3. Microsoft Outlook Express: POP3 and

IMAP4 Modes
For Outlook Express, the traffic generated depends on various parameters:

• Whether Outlook Express is configured to automatically check for new
messages at logon Outlook Express allows users to select whether or not to
connect to the server when starting.
• What quantity, size, and structure of the messages is to be downloaded The
quantity, size, and structure (the presence or absence of an attachment) of the
messages affect the traffic.
• Whether Outlook Today is displayed by default.
4.3.1. Test details

The following network traffic tests were performed on Outlook Express.
LOGON1: “Send and receive message at startup” option disabled
1. Create a POP3 or IMAP4 account in an empty mailbox.
2. Clear the Send and receive message at startup setting.
3. Run Outlook Express.
LOGON2: Inbox, No message, “Send and receive message at startup” option enabled
1. Create a POP3 or IMAP4 account empty mailbox.
2. Select the option Send and receive messages at startup.
3. Select the option When starting, go directly to my ‘Inbox’ folder.
LOGON3, LOGON4, and LOGON5: Inbox, 1–KB, 5–KB, and 10–KB messages, “Send
and receive message at startup” option enabled
2. Select the option Send and receive message at startup.
3. Select the option When starting, go directly to my Inbox folder.
4. Send 1–KB, 5–KB, and 10–KB messages.
LOGON6: Outlook Today, No message, “Send and receive message at startup” option
enabled
3. Clear the option When starting, go directly to my ‘Inbox’ folder.
LOGON7, LOGON8, LOGON9: Outlook Today, 1–KB, 5–KB, 10–KB messages, “Send
and receive message at startup” option enabled
3. Clear the option When starting, go directly to my ‘Inbox’ folder.
LOGOFF
1. Close all Outlook windows, except the main window.
2. Close Outlook by clicking File, then click Exit and Log Off.
4.3.2. Outlook Express IMAP4 Results

LOGON1 2115 23 807 12 1308 11
LOGON2 7065 81 2865 43 4200 38
LOGON6 4818 54 1957 28 2861 26
28 Client

LOGON3 15215 97 3626 48 11589 49
LOGON7 12025 68 2459 33 9566 35
LOGON4 42853 145 4674 63 38179 82
LOGON8 39657 118 3567 49 36090 69
LOGON5 77752 211 6179 85 71573 126
LOGON9 74853 191 5325 75 69528 116
LOGOFF 301 5 193 3 108 2
4.3.3. Outlook Express POP3 results


LOGON1 0 0 0 0 0 0
LOGON2
or 1172 16 511 8 661 8
LOGON6
LOGON3
or 7328 26 755 12 6573 14
LOGON7
LOGON4
or 31426 58 1491 24 29935 34
LOGON8
LOGON5
or 61611 99 2413 39 59198 60
LOGON9
LOGOFF 0 0 0 0 0 0
4.3.4. Outlook Express Measurements Analysis

When comparing an Outlook Express IMAP4 or POP3 client, keep in mind that
IMAP4 and POP3 are two transport protocols with different features. IMAP4 is
more powerful and advanced than POP3.
With an IMAP4 account, launching Outlook Express always generates 2 KB of
traffic to the server because it opens a connection with the Exchange 2000 server to
receive incoming message notification. With a POP3 account, Outlook Express
generates traffic only if a Send/Receive action is performed.
With IMAP4, displaying Outlook Today consumes less bandwidth than displaying
the Inbox. When Outlook Today is displayed, the client does not download headers
(sender, subject, received date). The number of messages and the contents affect
the traffic generated when displaying the Inbox folder at start up.
With POP3, there is no difference in network traffic whether Outlook Today is

displayed or not. POP3 always stores messages locally; Outlook Today reads the
messages stored on the hard drive. The only traffic in this situation is generated
during connection, and when messages are download.
IMAP4 uses 200 bytes to disconnect. In POP3, connection, download, and
disconnection are included in the Send/Receive process.
Overall, POP3 produces less network traffic than IMAP4, but does not offer the
same amount of functionality. POP3 can only download messages from one folder.
IMAP4 can synchronize many folders stored on a server. Outlook Express (with an
IMAP4 account) stores, by default, three folders on the server: Inbox, Drafts, and
Sent Items.
4.4. Netscape Messenger: POP3 & IMAP4 Modes
When a POP3 account is defined, Netscape Messenger does not offer an option to
download new messages at startup. Therefore, no traffic occurs. No capture
occurred in POP3 mode.
When an IMAP4 account is defined, Netscape Messenger automatically checks for
new messages at startup.
For Netscape Messenger, the generated network traffic depends on the quantity,
size, and structure of the messages to be downloaded.
4.4.1. Test Details

LOGON2: Inbox, no new message
1. Create an IMAP4 account in an empty mailbox.
2. Run Netscape Messenger.
LOGON3, LOGON4, LOGON5: Inbox, 1–KB, 5–KB, and 10–KB messages

2. Create an IMAP4 account in an empty mailbox for each test.
3. Run Netscape Messenger for each test.
LOGOFF
1. Close all Netscape windows except the main one.
2. Close Netscape Messenger by pressing ALT+F4.
30 Client
4.4.2. Netscape Messenger IMAP4 Results

Received by
Total Sent by Client
Client

LOGON2 2026 21 769 11 1257 10
LOGON3 3032 26 1153 14 1879 12
LOGON4 4776 28 1157 14 3619 14
LOGON5 7203 32 1277 16 5926 16
LOGOFF 785 9 355 5 430 4
4.4.3. Netscape Messenger IMAP4 measurements

Analysis
As there is no POP3 numbers for the reasons given above, only the numbers for IMAP4 are
available.
By default, Netscape Messenger does only store Inbox on the server. This fact can explain
the gap with Outlook Express logon.
The main advantage of Outlook Express in IMAP mode is that the server can notify it when
new messages are received; the user does not have to do it manually. Netscape Messenger
does not have this feature.
4.5. Outlook Web Access
For Outlook Web Access, the traffic generated depends on the quantity, size, and
structure of the messages to be downloaded.
4.5.1. Test Details

The following tests were performed:
LOGON2: Inbox, no new message
1. Create a mailbox.
2. Run Outlook Web Access (http://LONDON-04/exchange).
LOGON3, LOGON4, LOGON5: Inbox, 1–KB, 5–KB, and 10–KB new messages
LOGOFF
2. Close all Outlook Web Access windows.
3. Close Explorer.
4.5.2. Outlook Web Access Results


LOGON2 44597 75 10040 36 34557 39
LOGON3 45719 76 10623 37 35096 39
LOGON4 47751 77 10623 37 37128 40
LOGON5 50385 80 10677 38 39708 42
LOGOFF 54 1 54 1 0 0
4.5.3. Outlook Web Access Measurements Analysis

A quick analysis shows that Outlook Web Access generates more traffic than other
clients. Deeper analysis shows that part of the increased traffic comes from sending
the Web interface pages. The number of additional messages in the Inbox creates a
small amount of traffic (about 600 bytes per message).
Disconnecting from Exchange generates only 54 bytes of traffic, the least amount
of traffic among the tested messaging clients.
32 Client
5. Directory Access
Exchange 2000 introduces a new way to resolve address names. With the Active
Directory directory service, Exchange 2000 refers address query resolution to the
closest global catalog server.
The MAPI clients supporting this direct access method to the global catalog server
are:
• Microsoft Outlook 2000 (any version).
• Microsoft Outlook 98 version 8.5.6204.0 and later. This includes the Outlook 98
Archive Patch, available for download from the Microsoft Web site at
http://www.microsoft.com.
With earlier MAPI clients, Exchange 2000 forwards queries to the global catalog
server. Exchange 2000 server communicates with the global catalog server on
behalf the MAPI client.
Earlier MAPI clients are:
• Microsoft Outlook 98 versions before 8.5.6204.0.
• Microsoft Outlook 97 (any version).
• Microsoft Exchange client (any version).
Note Most companies that use Outlook 98 apply the Outlook 98

Archive Patch. With this update, Outlook 98 performs like
Outlook 2000. Therefore, measurements are included for
Outlook 97 97 8.04.5619 (from SR2b) and Outlook 2000.
These tests measured the network traffic generated when users check addresses or
names and other kinds of access information to the address book. Because the
clients differ, the tests are adapted to the four groups:
• MAPI clients: Microsoft Outlook 2000, Outlook 97
• Microsoft Outlook Express: POP3 and IMAP4 mode
• Netscape Messenger: POP3 and IMAP4 mode

• Web Client: Microsoft Outlook Web Access
The following concepts and their corresponding acronyms are referenced in this
section. The tests in this section are identified by the acronym that follows each
concept below.
Address Resolution (AR). When users check names or use automatic resolution
on a recipient.
Ambiguous Name Resolution (ANR). When a name is ambiguous, the user must
choose from a list of names. A Check Names action is performed before any ANR
tests to prevent excess traffic during initial access to the address book in the
Outlook session.
Address Lookup (AL). The address book dialog box appears when users click To
(or Cc, Bcc) in the Outlook client. The To button is clicked when addressing a
message to capture traffic resulting from users running MAPI queries of addresses
in the global address list (GAL).
Address Book View Lookup (ABVL). When users scroll through the GAL in the
address book.
Address Details (AD). When users ask for a GAL entry, or a name's properties.
5.2. MAPI Clients: Outlook 2000, Outlook 97
For MAPI clients, the traffic generated when accessing the directory depends on
various parameters:
• Whether or not the mailbox is new When a new Outlook client has connected
to a mailbox, does the first resolution generate extra traffic?
• Whether or not a profile is new When a profile has not yet been used, does the
first resolution generate extra traffic?
• How address resolution is accomplished Resolution can occur automatically or
by clicking Check Names.
• Whether an address is ambiguous. When the name is ambiguous, many results
return. How does this affect network traffic? A non-abiguous name or alias returns
only one result.
• Whether the name has been resolved prior to this instance. It is possible to
track a cache optimization.
34 Client
5.2.1. Test Details

The following network traffic tests were performed.
AR1: Check Names, first time in the profile (first Outlook session)
1. Create a new profile.
2. Start Outlook.
3. Create a new message.
4. Enter a non-ambiguous alias.
5. Click Check Names.
AR2: Check Names, existing profile, first Outlook session

1. Start Outlook.
AR3: Check Names, second Outlook session

This test preparation occurred after AR2.
2. Enter a non-ambiguous alias distinct from the alias entered in test AR2.
AR4: Check Names, name previously checked

This test preparation was performed after AR3.
2. Enter the non-ambiguous alias entered in test AR3.
AR5: Automatic name checking, first time in the profile (first Outlook session)
2. Start Outlook.
5. Click in Subject.
AR6: Automatic name checking, existing profile, first Outlook session

1. Start Outlook.
AR7: Automatic name checking, second Outlook session

2. Enter a non-ambiguous alias distinct from the alias entered in test AR6.
AR8: Automatic name checking, name previously checked

2. Enter the non-ambiguous alias entered in test AR7.
ANR1: Check Names, two results generated

1. Start Outlook.
3. Enter an ambiguous alias that will generate two results.
ANR2: Check Names, five names returned

1. Start Outlook.
3. Enter an ambiguous alias that will generate five results.
ANR3: Check Names, one of two names previously chosen

This test preparation was performed after ANR1.
1. Start Outlook.
3. Enter the ambiguous alias entered in test ANR1.
ANR4: Automatic name checking, two names returned

1. Start Outlook.
3. Enter an ambiguous alias with two results.
ANR5: Automatic name checking, five names returned

1. Start Outlook.
3. Enter an ambiguous alias that will generate five results.
36 Client
4. Click in the Subject box.
ANR6: Automatic resolution, one of two names previously chosen

This test preparation was performed after ANR4.
1. Start Outlook.
3. Enter the ambiguous alias entered in test ANR4.
AL1: To button, first time in the profile (first Outlook session)

2. Start Outlook.
4. Click To.
AL2: To button, first Outlook session

This test preparation was performed after AL1.
1. Start Outlook.
3. Click To.
AL3: To button, second Outlook session

This test preparation was performed after AL2.
2. Click To.
ABVL1: Scroll down one page in Address Book, first time in the profile
2. Start Outlook.
4. Click To.
5. Scroll down one page in the address list.
ABVL2: Scroll down one page in Address Book, first session

1. Start Outlook.
3. Click To.
4. Scroll down one page in the address list.
AD1: Check Properties on a name in the Address Book, first Outlook session
1. Start Outlook.
3. Click To.
4. Select an address.
5. Click Properties.
AD2: Double-click name in To line, first Outlook session

1. Start Outlook.
3. Enter the alias selected in test AD1.
5. Double-click name.
AD3: Check Properties on a name in the Address Book, second Outlook session
This test preparation was performed after AD1.
2. Click To.
3. Select an address.
AD4: Double-click name in To line, second Outlook session

This test preparation was performed after AD2.
2. Enter the alias selected in test AD1.
4. Double-click name.
Client to
Received by Exchange 2000
Total Sent by Client Client to GC Exchange 2000 GC to Client
Client Server to Client
Server
Frame
Bytes Frames Bytes Frames Bytes Frames Bytes Frames Bytes Bytes Frames Bytes Frames
s
AR1 1016 3 322 2 694 1 0 0 322 2 0 0 694 1
AR2 1016 3 322 2 694 1 322 2 0 0 694 1 0 0
AR3 1016 3 322 2 694 1 322 2 0 0 694 1 0 0
AR4 0 0 0 0 0 0 0 0 0 0 0 0 0 0
172
AR5 3460 11 1724 7 1736 4 0 0 7 0 0 1736 4
4
AR6 3460 11 1724 7 1736 4 856 5 866 2 1490 3 246 1
AR7 2348 8 858 5 1490 3 858 5 0 0 1490 3 0 0
AR8 1332 5 536 3 796 2 536 3 0 0 796 2 0 0
ANR1 1556 5 632 3 924 2 632 3 0 0 924 2 0 0
ANR2 2586 6 632 3 1954 3 632 3 0 0 1954 3 0 0
ANR3 1556 5 632 3 924 2 632 3 0 0 924 2 0 0
ANR4 2716 8 922 5 1794 3 922 3 0 0 1794 5 0 0
ANR5 4178 9 922 5 3256 4 922 5 0 0 3256 4 0 0
ANR6 1728 7 750 4 978 3 750 4 0 0 976 3 0 0
AL1 3844 5 306 2 3538 3 0 0 306 2 0 0 3538 3
AL2 3844 5 306 2 3538 3 306 2 0 0 3538 3 0 0
AL3 3844 5 306 2 3538 3 306 2 0 0 3538 3 0 0
ABVL1 3306 5 366 3 2940 2 0 0 366 3 0 0 2940 2
ABVL2 3246 4 306 2 2940 2 306 2 0 0 2940 2 0 0
AD1 4796 8 516 4 4280 4 516 4 0 0 4280 4 0 0
AD2 5556 12 944 6 4612 6 944 6 0 0 4612 6 0 0
AD3 1032 3 306 2 726 1 306 2 0 0 726 1 0 0
AD4 1792 7 724 4 1058 3 734 4 0 0 1058 3 0 0


AR1 1016 3 322 2 694 1
AR2 1016 3 322 2 694 1
AR3 1016 3 322 2 694 1
AR4 1016 3 322 2 694 1
AR5 4044 15 1992 9 2052 6
AR6 2435 9 874 5 1561 4
AR7 2380 8 874 5 1506 3
AR8 1364 5 552 3 812 2
ANR1 1556 5 632 3 924 2
ANR2 2586 6 632 3 1954 3
ANR3 1556 5 632 3 924 2
ANR4 2588 8 890 5 1698 3
ANR5 3984 9 890 5 3094 4
ANR6 1760 7 766 4 994 3
AL1 3904 6 366 3 3538 3
AL2 3904 6 366 3 3538 3
AL3 3904 6 366 3 3538 3
ABVL1 3306 5 366 3 2940 2
ABVL2 3306 5 366 3 2940 2
AD1 4796 8 516 4 4280 4
AD2 4796 8 516 4 4280 4
AD3 792 3 306 2 486 1
AD4 792 3 306 2 486 1
40 Client
5.2.4. Outlook 2000 and Outlook 97 Measurements

Analysis
Graph 2 : Address Resolution (comparative: Outlook 2000, Outlook 97)
4 500
4 000
3 500
3 000
2 500
Bytes
2 000
1 500
1 000
500
0 AR1 AR2 AR3 AR4 AR5 AR6 AR7 AR8
Outlook 2000 1 016 1 016 1 016 1 016 3 460 3 460 2 348 1 332
Outlook 97 1 016 1 016 1 016 1 016 4 044 2 435 2 380 1 364
AR1 = with button Check Names, 1st time in the profile (then 1st time in the session).
AR2 = with button Check Names, 1st time in that Outlook session.
AR3 = with button Check Names, 2nd time in that Outlook session.
AR4 = with button Check Names, name previously checked.
AR5 = automatic resolution, 1st time in the profile (then 1st time in the session).
AR6 = automatic resolution, 1st time in that Outlook session.
AR7 = automatic resolution, 2nd time in that Outlook session.
AR8 = automatic resolution, name previously checked.
Graph 3: Ambiguous Name Resolution (comparative: Outlook 2000, Outlook 97)
4 500
4 000
3 500
3 000
2 500
Bytes
2 000
1 500
1 000
500
0 ANR1 ANR2 ANR3 ANR4 ANR5 ANR6
Out look 2000 1 556 2 586 1 556 2 716 4 178 1 728

Out look 97 1 556 2 586 1 556 2 588 3 954 1 760
ANR1 = with button Check Names, 2 names returned.

ANR2 = with button Check Names, 5 names returned.
ANR3 = with button Check Names, one of 2 names previously checked.
ANR4 = automatic resolution, 2 names returned.
ANR5 = automatic resolution, 5 names returned.
ANR6 = automatic resolution, one of 2 names previously checked.
Graph 4 : Address Lookup, Address Book View Lookup, Address Details (comparative:
Outlook 2000, Outlook 97)
6 000
5 000
4 000
Bytes
3 000
2 000
1 000
0 AL1 AL2 AL3 ABVL1 ABVL2 AD1 AD2 AD3 AD4
Out look 2000 3 844 3 844 3 844 3 306 3 246 4 796 5 556 1 032 1 792
Out look 97 3 904 3 904 3 904 3 306 3 306 4 796 4 796 792 792
AL1 = with button To, 1st time in the profile (then 1st time in the session).
AL2 = with button To, 1st time in that Outlook session.
AL3 = with button To, 2nd time in that Outlook session.
ABVL1 = scroll down once, 1st time in the profile.
ABVL2 = scroll down once, 1st time in the session.
AD1 = double-click 1st time in Outlook session on a resolved address in Address Book.
AD2 = double-click 1st time in Outlook session on a resolved address in To Line.
AD3 = double-click 2nd time in Outlook session on a resolved address in Address Book
AD4 = double-click 2nd time in Outlook session on a resolved address in To Line.
A first analysis shows that Outlook 2000 queries the global catalog server only to
resolve names. Only upon the second launch of Outlook 2000 (using the profile) is
the global catalog server queried. When a new profile is used, the GAL name is not
known, so Outlook cannot resolve it. It is during profile initialization (the first log
on) that Exchange 2000 gives the client the GAL name. Outlook 2000 stores this
name in its profile in a registry entry:
HKEY_CURRENT_USER \ Software \ Microsoft \ Windows NT \

CurrentVersion \ Windows Messaging Subsystem \ Profiles \ profile
name \ dca740…2fe182
Value name: 001e6602
Value type: String
Value data: \\DirectoryServer.domain
42 Client
Upon the second launch, every initial name resolution (manual or automatic) in the
Outlook 2000 session generates traffic with the Exchange server and the global
catalog server. Subsequent queries are sent directly to the global catalog server. The
Exchange 2000 server does not process any more name queries.
On the other hand, Outlook 97 communicates only with the Exchange 2000 server,
which functions like a proxy server between the global catalog server and
Outlook 97.
Outlook 2000 caches all resolved names during the session. This cache is only
written or read when users click Check Names (or CTRL+K). The automatic
resolution does not use cache; it uses the global catalog. The automatic resolution
always contacts the Exchange 2000 server and the global catalog server during the
first resolution in the session.
The ambiguous-name resolution generates the same amount of traffic on
Outlook 97 and Outlook 2000. ANR always generates the same levels of traffic
with Check Names, but it generates less traffic when using automatic resolution on
a resolved name.
The traffic generated when the global address list is displayed in the address book
does not depend on the number of entries. The initial traffic only depends on the
visible entries on the first page.
In Outlook 2000, viewing address details from the To box generates 70 percent
more bytes than viewing details from the address book. In Outlook 97, both
methods generate the same amount of traffic. An initialization phase occurs when
viewing details during the first Outlook session, because the details form design
downloads. Showing other tabs during the first Outlook session also generates
extra traffic.
Generally, both versions of Outlook generate about the same amount of traffic. The
difference is in how the traffic is spread out. Outlook 2000 directly queries the
global catalog server for all address book-related actions; this consumes fewer
CPU cycles on the Exchange 2000 server.
5.3. Outlook Express: LDAP mode
Outlook Express can resolve names with LDAP. Because LDAP is a protocol used
for accessing directories, it is independent of the mailbox access protocols. You
will need to create an LDAP account. It is not necessary to perform tests with
POP3 and IMAP4. There is no difference between Check Names and Find
Address; the same LDAP query occurs.
5.3.1. Test Details

AR1: Find people
1. Open the Find People form.
3. Click Find Now.
ANR1: Find people, two names returned

1. Show the Find People form.
2. Enter an ambiguous alias that generates two results.
3. Click Find Now.
ANR2: Find people, five names returned

1. Open the Find People form.
2. Enter an ambiguous alias that generates five results.
3. Click Find Now.
5.3.2. Outlook Express LDAP Results

Received by
Client

AR1 8292 34 4978 19 3314 15
ANR1 12348 60 8992 34 3356 26
ANR2 13524 60 8992 34 4532 26
Analysis
These measurements allow us to quantify the traffic for find one person to 8 KB
per non-ambiguous name. Finding an ambiguous name generates logically more
traffic in bytes term.
5.4. Netscape Messenger: LDAP mode
An unexpected authentication issue prevented completion of LDAP tests with

Netscape Messenger.
44 Client
The only way to resolve names in Outlook Web Access is to click Check Names
(CTRL+K is allowed). The automatic resolution functionality does not exist in
Outlook Web Access. Outlook Web Access does not perform resolved names
caching, so many of the tests for other clients are not useful because for Outlook
Web Access the tests always produce the same result. Outlook Web Access does
not provide addressbook lookup; you must find the name you are looking for. The
To and Cc and Bcc buttons display the same UI to find people.
5.5.1. Test Details

AR1: Check Names
2. In the To box, enter a non-ambiguous alias.
ANR1: Check Names, two names returned

2. In the To box, enter an ambiguous alias that will generate two results.
3. Click Check Names, and then choose a name.
ANR2: Check Names, five names returned

2. In the To box, enter an ambiguous alias that will generate five results.
3. Click Check Names, and then choose a name.
AL1: To button (only show Find Names window), second launch

2. Click To.
AD1: To button, double-click name for details then Cancel

4. Double-click the resolved name then Cancel.
AD2: To button, double-click name for details then OK

4. Double-click the resolved name then OK.

Received by
Client

AR1 7507 17 4697 10 2810 7
ANR1 19150 40 10289 22 8861 18
ANR2 19485 40 10289 22 9196 18
AL1 5149 7 453 3 4696 4
AD1 3973 7 458 3 3515 4
AD2 11452 21 5175 11 6277 10
5.5.3. Outlook Web Access Measurements Analysis

Most of the Outlook Web Access traffic comes from the HTML page description
and the XML content. The address-resolution process downloads only the XML
content; this dramatically reduces the traffic, because the entire page is not
downloaded. The Check Names traffic is linked to the amount of text in the body:
when you check names the names are underlined then the UI is refreshed.
Therefore, it is recommended to check names before entering text in body or to not
check names if the name is not ambiguous. To reduce traffic, it is recommended to
cancel every window with settings if you did not any changes (example: E-mail
properties).
46 Client
6. Generic Mail Item Tests
These generic mail-item tests include the more common tasks performed by a
messaging client: sending, reading, modifying, and deleting messages, and opening
attachments.
6.1. Tests performed
The following list describes the general tasks that each e-mail client performed for
the tests.
• Send a 1 KB, 2 KB, and 4 KB item Send a mail message containing either
1 KB, 2 KB, or 4 KB of text. Include a simple subject indicating the contents of
the message. The recipient for the message is the current logged-on user. Send
messages as:
• RTF
• HTML
• Plain Text
• Send an item with a 10 KB, 50 KB, 100 KB, 500 KB, or 1,000 KB
attachment Send a mail message containing a 10 KB, 50 KB, 100 KB, 500 KB,
or 1,000 KB attachment, with no other text. Include a simple subject indicating the
contents of the message, such as “1 KB Attachment.” Send messages as:
• RTF
• HTML
• Plain Text
• Read an item: x KB (with or without attachment) Open the items sent in

previous test.
• Open the attachment within items Open the attachments in the items sent in
previous test.
• Delete an item x KB (with or without attachment) Delete the items sent in
previous test.
• Delete y items of xKB item Empty the deleted items folder, which contains
y number of x–KB sized items.
• Process read receipt Process a 1–KB mail message sent with a read receipt. The
user double-clicks on the read receipt.
• Process delivery receipt Processed a 1–KB mail message sent with a read
receipt. The user double-clicks on the read receipt.
• Create subfolder Create a single subfolder in the Inbox folder.
6.2. Test Details

MSG1: New message, first Outlook session
1. Start Outlook.
2. Create a new mail message.
MSG2: New message, second Outlook session

SR1, SR2, SR3: Send an RTF message with 1 KB, 2 KB, or 4 KB of text
1. Open the x KB .msg file where x is 1 KB, 2 KB, or 4 KB.
2. Copy the text.
3. Create a new RTF mail message.
4. In the To box, enter a recipient other than yourself.
5. In the Subject box, enter a brief description of the test.
6. Paste the text in the message body.
7. Click Send.
SR4, SR5, SR6, SR7, SR8: Send an RTF message with 1 KB of text and an attachment
of 10 KB, 50 KB, 100 KB, 500 KB, or 1,000 KB.
1. Open the 1–KB .msg file.
2. Copy the text.
3. Create an RTF mail message.
5. In the Subject box, type R1–KBx where x is 0050, 0100, 0500, or 1000
(corresponding to the size of the attachment).
7. Insert the corresponding attachment.
8. Click Send.
SH1, SH2, SH3: Send an HTML message with 1 KB, 2 KB, or 4 KB of text
1. Open the x KB .msg file where x is 1 KB, 2 KB, or 4 KB.
2. Copy the text.
48 Client
3. Create a new HTML mail message.

5. In the Subject box, enter Hx KB0000 where x is 1 KB, 2 KB, or 4 KB.
7. Click Send.
SH4, SH5, SH6, SH7, SH8: Send an HTML message with 1 KB of text and an
attachment of 10 KB, 50 KB, 100 KB, 500 KB, or 1,000 KB.
1. Open the 1–KB .msg file.
2. Copy the text.
5. In the Subject box, type H1 KBx where x is 0010, 0050, 0100, 0500, or
1000 (corresponding to the size of the attachment).
8. Click Send.
SP1, SP2, SP3: Send a Plain Text message with 1 KB, 2 KB, or 4 KB of text.
1. Open the x–KB .msg file where x is 1 KB, 2 KB, or 4 KB.
2. Copy the text.
3. Create a new Plain Text mail message.
5. In the Subject box, type Px KB0000 where x is 1 KB, 2 KB, or 4 KB.
7. Click Send.
SP4, SP5, SP6, SP7, SP8: Send a Plain Text message with 1 KB of text and an
attachment of 10 KB, 50 KB, 100 KB, 500 KB, or 1,000 KB.
1. Open the 1KB .msg file.
2. Copy the text.
3. Create a new Plain Text mail message.
5. In the Subject box, type P1 KBx where x is 0050, 0100, 0500, or 1000
(corresponding to the size of the attachment).
8. Click Send.
RR1, RR2, RR3: Open an RTF message with 1 KB, 2 KB, or 4 KB of text
1. Open the RTF message Rx KB0000 where x is 1 KB, 2 KB, or 4 KB.
RR4: Open an RTF message with 1 KB of text and an attachment of 10 KB.

1. Open the RTF message R1 KB0010 and open the attachment.
RH1, RH2, RH3: Open an HTML message with 1 KB, 2 KB, or 4 KB of text
1. Open the HTML message Hx KB0000 where x is 1 KB, 2 KB, or 4 KB.
RH4: Open an HTML message with 1 KB of text and an attachment of 10 KB.

1. Open the HTML message H1 KB0010 and open the attachment.
RP1, RP2, RP3: Open a Plain Text message with 1 KB, 2 KB, or 4 KB of text
1. Open the Plain Text message Px KB0000 where x is 1 KB, 2 KB, or 4 KB.
RP4: Open a Plain Text message with 1 KB of text and an attachment of 10 KB.
1. Open the Plain Text message P1 KB0010 and open the attachment.
OR4, OR5, OR6, OR7, OR8: Open the 10–KB, 50–KB, 100–KB, 500–KB, and 1,000–KB
attachments in the 1–KB RTF messages.
1. Open the RTF message R1 KB0010 without opening the attachment.
2. Open the attachment.
OH4, OH5, OH6, OH7, OH8: Open the 10–KB, 50–KB, 100–KB, 500–KB, and 1,000–KB
attachment in the 1–KB HTML message.
1. Open the HTML message H1 KB0010 without opening the attachment.
OP4, OP5, OP6, OP7, OP8: Open the 10–KB, 50–KB, 100–KB, 500–KB, and 1,000–KB
attachments in the 1–KB Plain Text messages.
1. Open the Plain Text message, P1 KB0010, without opening the attachment.
D1: Delete a message with no attachments

1. Select a message with no attachments.
2. Press Delete.
D2: Delete a message with one attachment

1. Select a message with one attachment.
2. Press Delete.
PRR1: Open a read receipt.

1. Open a read receipt.
50 Client
PDR1: Open a delivery receipt

1. Open a delivery receipt.
DI1: Empty Deleted Items containing a 1–KB message

1. Put one 1–KB message in the Deleted Items folder.
2. Empty the Deleted Items folder.
DI2: Empty Deleted Items containing one 1–KB message with one 1,000–KB
attachment
1. Put one 1–KB message containing one 1,000–KB attachment in the Deleted
Items folder.
DI3: Empty Deleted Items containing ten 1KB messages

1. Put ten 1–KB messages in the Deleted Items folder.
DI4: Empty Deleted Items containing ten 1–KB messages with one 1,000–KB
attachment each
1. Put ten 1–KB messages each containing a 1,000–KB attachment in the
Deleted Items folder.
FOL1: Create one folder

1. Open Inbox
2. Create a new folder.
6.3. Outlook 2000 Results

Create Message (1st
time in that Outlook 1688 3 1410 2 278 1
session)
Create Message (2nd
time in that Outlook 0 0 0 0 0 0
session)
Send an Item 1K RTF 5354 18 2798 10 2556 8

Send an Item 1K
15794 28 12652 16 3142 12
HTML
Send an Item 1K Plain
5958 20 2996 11 2962 9
Text
Send an Item 2K RTF 6442 18 3854 10 2588 8
Send an Item 2K
27822 41 24380 24 3442 17
HTML
6950 20 3988 11 2962 9
Text
Send an Item 4K RTF 8386 21 5738 12 2648 9
Send an Item 4K
50594 64 46594 40 4000 24
HTML
9018 24 5996 14 3022 10
Text
Send an Item with
21192 40 17534 24 3658 16
Attachment 10K RTF
Send an Item with
Attachment 10K 31104 49 27266 30 3838 19
HTML
Send an Item with
Attachment 10K Plain 21120 38 17462 22 3658 16
Text
Send an Item with
65272 83 60788 51 4484 32
Attachment 50K RTF
Send an Item with
Attachment 50K 75244 93 70580 58 4664 35
HTML
Send an Item with
Attachment 50K Plain 65316 84 60848 52 4468 32
Text
Send an Item with
120792 143 114730 90 6062 53
Attachment 100K RTF
Send an Item with
Attachment 100K 130644 151 124402 95 6242 56
HTML
Send an Item with
Attachment 100K 120748 142 114686 89 6062 53
Plain Text
Send an Item with
564422 616 545210 388 19212 228
Attachment 500K RTF
Send an Item with
Attachment 500K 574394 626 555002 395 19392 231
HTML
Send an Item with
Attachment 500K 564498 617 545286 389 19212 228
Plain Text
Send an Item with
Attachment 1000K 1118554 1210 1082916 763 35638 447
RTF
52 Client

Send an Item with
Attachment 1000K 1128466 1219 1092648 769 35818 450
HTML
Send an Item with
Attachment 1000K 1118630 1211 1082992 764 35638 447
Plain Text
Read an Item 1K RTF 5084 8 1108 4 3976 4
Read an Item 1K
17212 21 2292 9 14920 12
HTML
Read an Item 1K Plain
5116 8 1108 4 4008 4
Text
Read an Item 2K RTF 6060 8 1108 4 4952 4
Read an Item 2K
28924 29 2292 9 26632 20
HTML
6092 8 1108 4 4984 4
Text
Read an Item 4K RTF 8346 12 1334 6 7012 6
Read an Item 4K
68714 86 5396 37 63318 49
HTML
9974 14 2124 7 7850 7
Text
Read an Item with
10056 17 1778 8 8278 9
Attachment RTF
Read an Item with
22048 28 2826 11 19222 17
Attachment HTML
Read an Item with
10088 17 1778 8 8310 9
Attachment Plain Text
Open the 10K
11892 19 962 8 10930 11
Attachment
Open the 50K
56834 59 2012 18 54822 41
Attachment
Open the 100K
112252 108 3272 30 108980 78
Attachment
Open the 500K
556552 508 13772 130 542780 378
Attachment
Open the 1000K
1111278 1007 26792 254 1084486 753
Attachment
Delete an Item 1680 7 558 4 1122 3
Delete an Item with
1712 7 574 4 1138 3
Attachment
Process read receipt 12412 31 4050 16 8362 15
Process delivery
13990 30 3868 15 10122 15
receipt
Delete1 item of 1K
item from “Deleted 1520 7 558 4 962 3
Items Folder”

Delete 1 items with
1000K Attachment
1520 7 558 4 962 3
from “Deleted Items
Folder”
Delete10 item of 1K
item from “Deleted 1856 7 622 4 1234 3
Items Folder”
Delete 10 items with
1000K Attachment
1856 7 622 4 1234 3
from “Deleted Items
Folder”
Create folder 2852 13 1200 7 1652 6
With Outlook 2000, an attachment is loaded onto the server as soon as it is inserted
in the message. When an attachment is inserted in a message before the user
composes the message, the attachment has more time to load on the server and,
when the user clicks Send, the time before the message is sent is reduced.
This functionality is dependant attachment size. A 50–KB attachment immediately
uploads during composition, while 10–KB attachments upload only when the user
clicks Send.
The amount of traffic generated when a new folder is created depends on the
number of folders in the mailbox. In these tests, the capture occurred with standard
Outlook folders. Every new folder adds about 100 bytes in generated traffic.
The deletion (tests D1, D2) of items moves the items to the Deleted Items folder.
Graph 5 : Generic Mail Item / Outlook 2000 - Send item without attachment
(comparative: RTF, HTML, Plain Text)
60 000
50 000
40 000
Bytes
30 000
20 000
10 000
0 1K 2K 4K
RTF 5 354 6 442 8 386

HTM L 15 794 27 822 50 594
Plain Text 5 958 6 950 9 018
54 Client
HTML messages consume more bandwidth than standard Outlook messages.

However, the amount of bandwidth consumed is dependent on the content of the
message. The tests on public folders later in this section show that message form
can dramatically reduce or increase the size difference between HTML format,
RTF, or Plain Text. HTML is the most voluminous message format, and RTF is the
lightest format in Outlook 2000. Outlook 2000 does not download the attachment,
if there is one, when opening a message.
Graph 6 : Generic Mail Item / Outlook 2000 - Send a 1K item with attachment
(comparative: RTF; HTML, Plain Text)
1 200 000
1 000 000
800 000
Bytes
600 000
400 000
200 000
0
1K t ext w / 10 K 1K t ext w / 1K t ext w / 1K t ext w / 1K t ext w /
O f f ice 50 K O f f ice 10 0 K Of f ice 50 0 K O f f ice 10 0 0 K O f f ice
at t achment at t achment at t achment at t achment at t achment
RTF 2 1 19 2 6 5 2 72 12 0 79 2 56 4 4 2 2 1 118 554

HT M L 3 1 10 4 75 2 4 4 13 0 6 4 4 574 3 9 4 1 12 8 4 6 6
Plain T ext 2 1 12 0 6 5 3 16 12 0 74 8 56 4 4 9 8 1 118 6 3 0
The size difference between message types is mainly due to attachment size. There
is no preferred format in which to send an attachment, because attachment
conversion does not affect the message size: 1 MB stays at about 1 MB after
conversion. The size difference between formats is due to the 1 KB of converted
text.
Graph 7 : Generic Mail Item / Outlook 2000 - Read item without attachment
70 000
60 000
50 000
40 000
Bytes
30 000
20 000
10 000
0
1K message with no 2K message with no 4K message with no
attachment attachment attachment
RTF 5 084 6 060 8 346

HTM L 17 212 28 924 68 714
Plain Text 5 116 6 092 9 974
The analysis is the same than the one for sending messages: HTML is the heaviest
format, RTF is the lightest one.
Graph 8: Generic Mail Item / Outlook 2000 - Open and Read a 1K message with an
attachment (comparative: RTF, HTML, Plain Text)
1 200 000
1 000 000
800 000
Bytes
600 000
400 000
200 000
0 1K t ext w/
1K t ext w/ 10K 1K t ext w/ 50K 1K t ext w/ 100K 1K t ext w/ 500K
1000K
RTF 21 948 66 890 122 308 566 608 1 121 334

HTM L 33 940 78 882 134 300 578 600 1 133 326
Plain Text 21 980 66 922 122 340 566 640 1 121 366
The analysis is the same than the one for sending messages with attachments: no
preferred format. As Outlook 2000 does not download the attachment when
opening such message, these tests include opening attachments. Otherwise, we
would have the same results whatever the attachment size.
56 Client

Create Message 1608 3 1330 2 278 1
(1st time in that
Outlook session)
Create Message 0 0 0 0 0 0
(2nd time in that
Outlook session)
Send an Item 1K 5358 17 2834 9 2524 8
Send an Item 2K 6394 18 3870 10 2524 8
Send an Item 4K 8418 21 5834 12 2584 9
Send an Item with 35106 37 31976 22 3130 15
Attachment 10K
Send an Item with 85004 104 80116 65 4888 39
Attachment 50K
Send an Item with 120352 140 114488 88 5864 52
Attachment 100K
Send an Item with 564102 615 545088 388 19014 227
Attachment 500K
Send an Item with 1118234 1209 1082794 763 35440 446
Attachment 1000K
Read an Item 1K 5440 12 1222 6 4218 6
Read an Item 2K 6508 13 1298 7 5210 6
Read an Item 4K 8734 16 1464 8 7270 8
Read an Item with 10048 19 1726 9 8322 10
Attachment
Open the 10K 15358 24 1248 11 14110 13
Attachment
Open the 50K 58286 73 2898 31 55388 42
Attachment
Open the 100K 114424 134 4878 55 109546 79
Attachment
Open the 500K 564724 634 21378 255 543346 379
Attachment
Open the 1000K 1127010 1259 41958 505 1085052 754
Attachment
Delete an Item 1784 9 602 5 1182 4
Delete an Item 1800 9 602 5 1198 4
with Attachment
Process read 17318 41 5426 21 11892 20
receipt

Process delivery 16860 38 6640 20 10220 18
receipt
Delete1 item of 1K 1932 9 756 5 1176 4
item from
“Deleted Items
Folder”
Delete 1 items 1932 9 756 5 1176 4
with 1000K
Attachment from
“Deleted Items
Folder”
Delete10 item of 2492 9 756 5 1736 4
1K item from
“Deleted Items
Folder”
Delete 10 items 2492 9 756 5 1736 4
with 1000K
Attachment from
“Deleted Items
Folder”
Create subfolder 2360 10 848 6 1512 4
Outlook 97 can only send RTF messages. The Deleting an item option means that
messages are moved to the Deleted Items folder. The act of emptying the Deleted
Items folder makes no impact on network traffic, whether or not messages have
attachments.
6.5. Outlook Express IMAP Results

Refresh Inbox 1172 16 511 8 661 8
with no new
message
Send an Item 1K 34344 72 31659 43 2685 29
HTML
Send an Item 1K 16780 62 14281 34 2499 28
Plain Text
Send an Item 2K 67005 114 63426 68 3579 46
HTML
58 Client

Send an Item 2K 30239 84 27500 42 2739 42
Plain Text
Send an Item 4K 127707 176 122928 110 4779 66
HTML
Send an Item 4K 55383 94 52303 56 3080 38
Plain Text
Send an Item with 63793 110 60334 66 3459 44
Attachment 10K
HTML
Send an Item with 45710 89 42725 52 2985 37
Attachment 10K
Plain Text
Send an Item with 184331 236 178352 150 5979 86
Attachment 50K
HTML
Send an Item with 166267 214 160768 136 5499 78
Attachment 50K
Plain Text
Send an Item with 332972 388 323993 252 8979 136
Attachment 100K
HTML
Send an Item with 315138 370 306519 240 8619 130
Attachment 100K
Plain Text
Send an Item with 1523040 1620 1489281 1072 33759 549
Attachment 500K
HTML
Send an Item with 1505086 1601 1471807 1060 33279 541
Attachment 500K
Plain Text
Send an Item with 3009231 4196 2944572 2098 64659 2098
Attachment 1000K
HTML
Send an Item with 2991219 3141 2927100 2086 64119 1055
Attachment 1000K
Plain Text
Read an Item 1K 16800 30 738 11 16062 19
HTML
Read an Item 1K 7766 20 558 8 7208 12
Plain Text
Read an Item 2K 32711 43 978 15 31733 28
HTML
Read an Item 2K 14556 27 738 11 13818 16
Plain Text
Read an Item 4K 62882 71 1398 22 61484 49
HTML
Read an Item 4K 27150 37 858 13 26292 24
Plain Text

Read an Item with 31220 43 978 15 30242 28
Attachment 10K
HTML
Read an Item with 22355 35 858 13 21497 22
Attachment 10K
Plain Text
Read an Item with 90973 97 1938 31 89035 66
Attachment 50K
HTML
Read an Item with 82107 89 1758 28 80349 61
Attachment 50K
Plain Text
Read an Item with 164887 166 3138 51 161749 115
Attachment 100K
HTML
Read an Item with 155902 156 2898 47 153004 109
Attachment 100K
Plain Text
Read an Item with 755728 710 12738 211 742990 499
Attachment 500K
HTML
Read an Item with 746802 701 12558 208 734244 493
Attachment 500K
Plain Text
Read an Item with 1493539 1389 24738 411 1468801 978
Attachment 1000K
HTML
Read an Item with 1484499 1378 24498 407 1460001 971
Attachment 1000K
Plain Text
Delete an Item 544 7 290 4 254 3
Process read 3036 21 678 10 2358 11
receipt
Delete1 item of 1K 534 7 253 4 281 3
item (Purge
Deleted Messages)
Delete 1 items 534 7 253 4 281 3
with 1000K
Attachment (Purge
Deleted Messages)
Delete10 item of 651 7 253 4 398 3
1K item (Purge
Deleted Messages)
Delete 10 items 651 7 253 4 398 3
with 1000K
Attachment (Purge
Deleted Messages)
60 Client
By default, the Save copy of sent message in the Sent Items Folder option is
selected. This option causes the message to be sent twice. When you clear it, you
can divide all sending operations by two.
When opening a message with attachments, the message and the attachments are
stored locally. If they are both read later, there is no more traffic. Netscape
Messenger generates traffic each time you want to read the attachment.
As soon as a message is received on the server, its header is sent to the client.
There is no need to generate a refresh.
Graph 9 : Generic Mail Item / Outlook Express / IMAP - Send item without attachment
(comparative: HTML, Plain Text)
140 000
120 000
100 000
80 000
Bytes
60 000
40 000
20 000
0 1K message 2K message 4K message
HTM L 34 344 67 005 127 707

Plain Text 16 780 30 239 55 383
The conclusions are the same as those with Outlook 2000. There is a big difference
in message size between a message sent in HTLM format and a message sent in
Plain Text.
Graph 10 : Generic Mail Item / Outlook Express / IMAP - Send item with attachment
3 500 000
3 000 000
2 500 000
2 000 000
Bytes
1 500 000
1 000 000
500 000
0 1K text w/ 10K 1K t ext w/ 50K 1K t ext w/ 100K 1K text w/ 500K 1K t ext w/

Off ice Off ice Off ice Of fice 1000K Of f ice
HTM L 63 793 184 331 332 972 1 523 040 3 009 231
Plain Text 45 710 166 267 315 138 1 505 086 2 991 219
The conclusions are the same ones done with Outlook 2000. The small gap
between formats is about the text format itself. There is no correlation between the
attachment’s size and the gap.
Graph 11 : Generic Mail Item / Outlook Express / IMAP - Read Item without
attachment (comparative: HTML, Plain Text)
70 000
60 000
50 000
40 000
Bytes
30 000
20 000
10 000
HTM L 16 800 32 711 62 882

Plain Text 7 766 14 556 27 150
Same analysis than the one for sending messages with none attachment.
62 Client
Graph 12 : Generic Mail Item / Outlook Express / IMAP - Read item with attachment
1 600 000
1 400 000
1 200 000
1 000 000
Bytes
800 000
600 000
400 000
200 000
HTM L 31 220 90 973 164 887 755 728 1 493 539

Plain Text 22 355 82 107 155 902 746 802 1 484 499
Same analysis than the one for sending messages with none attachment.
6.6. Outlook Express POP Results

Refresh with no 1172 16 511 8 661 8
new message
Send an Item 1K 17149 37 15803 21 1346 16
HTML
Send an Item 1K 9344 44 7517 23 1827 21
Plain Text
Send an Item 2K 34403 69 32096 40 2307 29
HTML
Send an Item 2K 16073 50 14126 27 1947 23
Plain Text
Send an Item 4K 64814 101 61847 61 2967 40
HTML
Send an Item 4K 28841 63 26654 36 2187 27
Plain Text
Send an Item with 32857 68 30550 39 2307 29
Attachment 10K
HTML

Send an Item with 23931 59 21804 33 2127 26
Attachment 10K
Plain Text
Send an Item with 93126 131 89559 81 3567 50
Attachment 50K
HTML
Send an Item with 84086 120 80759 74 3327 46
Attachment 50K
Plain Text
Send an Item with 166506 208 162379 132 4127 76
Attachment 100K
HTML
Send an Item with 158580 199 153633 126 4947 73
Attachment 100K
Plain Text
Send an Item with 762390 822 745023 542 17367 280
Attachment 500K
HTML
Send an Item with 753414 813 736227 536 17187 277
Attachment 500K
Plain Text
Send an Item with 1505455 1592 1472668 1055 32787 537
Attachment 1000K
HTML
Send an Item with 1496529 1583 1463922 1049 32607 534
Attachment 1000K
Plain Text
Read an Item 1K 17008 34 807 13 16201 21
HTML
Read an Item 1K 8685 31 957 14 7728 17
Plain Text
Read an Item 2K 33513 52 1257 19 32256 33
HTML
Read an Item 2K 15357 36 1017 15 14340 21
Plain Text
Read an Item 4K 63638 79 1677 26 61961 53
HTML
Read an Item 4K 28011 47 1197 18 26814 29
Plain Text
Read an Item with 31962 51 1257 19 30705 32
Attachment 10K
HTML
Read an Item with 23096 43 1137 17 21959 26
Attachment 10K
Plain Text
Read an Item with 91600 103 2037 32 89563 71
Attachment 50K
HTML
64 Client

Read an Item with 82794 96 1977 31 80817 65
Attachment 50K
Plain Text
Read an Item with 165341 169 3117 50 162224 119
Attachment 100K
HTML
Read an Item with 156715 165 3237 52 153478 113
Attachment 100K
Plain Text
Read an Item with 754201 680 10737 177 743464 503
Attachment 500K
HTML
Read an Item with 745275 671 10557 174 734718 497
Attachment 500K
Plain Text
Read an Item with 1494762 1402 25227 417 1469535 985
Attachment 1000K
HTML
Read an Item with 1485356 1388 24837 412 1460519 976
Attachment 1000K
Plain Text
Process read 3731 28 957 14 2774 14
receipt
There is no refresh test in POP3 mode; the amount of network traffic is the same as
that for reading messages.
Graph 13 : Generic Mail Item / Outlook Express 5.01 / POP - Send item without
70 000
60 000
50 000
40 000
Bytes
30 000
20 000
10 000
HTM L 17 149 34 403 64 814

Plain Text 9 344 16 073 28 841
HTML format consumes more bandwidth than Plain Text, even if the text has no
rich-text formatting.
Graph 14 : Generic Mail Item / Outlook Express / POP - Send item with attachment
1 600 000
1 400 000
1 200 000
1 000 000
Bytes
800 000
600 000
400 000
200 000
0 1K text w/ 10K 1K text w/ 50K 1K text w/ 100K 1K text w/ 500K 1K text w/
Office Off ice Of fice Off ice 1000K Of fice
HTM L 32 857 93 126 166 506 762 390 1 505 455

Plain Text 23 931 84 086 158 580 753 414 1 496 529
The difference between POP and HTML formats is mainly caused by the size of
attachments. There is no preferred format in which to send an attachment.
Attachment conversion does not change the amount of network traffic: 1 MB stays
at about 1 MB after conversion. The size difference between formats is due to the
1 KB of converted text.
Graph 15 : Generic Mail Item / Outlook Express 5.01 / POP - Read Item without
70 000
60 000
50 000
40 000
Bytes
30 000
20 000
10 000
HTM L 17 008 33 513 63 638

Plain Text 8 685 15 357 28 011
66 Client
Graph 16 : Generic Mail Item / Outlook Express 5.01 / POP - Read Item with
1 600 000
1 400 000
1 200 000
1 000 000
Bytes
800 000
600 000
400 000
200 000
0 1K t ext w/ 10K 1K t ext w/ 50K 1K t ext w/ 100K 1K text w/ 500K 1K text w/
Off ice Off ice Off ice Off ice 1000K Off ice
HTM L 31 962 91 600 165 341 754 201 1 494 762

Plain Text 23 096 82 794 156 715 745 275 1 485 356
6.7. Netscape Messenger IMAP Results

Refresh with no new 200 3 123 2 77 1
message
Refresh with 1 new 1129 7 449 4 680 3
message
Refresh with 5 new 3047 11 513 5 2534 6
message
Refresh with 10 new 5247 13 573 6 4707 7
message
Send an Item 1K 13669 41 12050 24 1619 17
HTML
Send an Item 1K 13630 40 11983 23 1647 17
Plain Text
Send an Item 2K 25740 56 23797 34 1943 22
HTML
Send an Item 2K 26089 59 23992 36 2097 23
Plain Text
Send an Item 4K 48418 84 45935 53 2483 31
HTML

Send an Item 4K 48463 85 45916 53 2547 32
Plain Text
Send an Item with 43162 77 40799 48 2363 29
Attachment 10K
HTML
Send an Item with 43219 78 40852 49 2367 29
Attachment 10K
Plain Text
Send an Item with 165118 225 159815 147 5303 78
Attachment 50K
HTML
Send an Item with 165115 225 159808 147 5307 78
Attachment 50K
Plain Text
Send an Item with 315643 407 306679 268 8964 139
Attachment 100K
HTML
Send an Item with 315698 408 306671 268 9027 140
Attachment 100K
Plain Text
Send an Item with 1520309 1872 1481885 1242 38424 630
Attachment 500K
HTML
Send an Item with 1520064 1868 1481877 1242 38187 626
Attachment 500K
Plain Text
Send an Item with 3024199 3694 2949352 2457 74847 1237
Attachment 1000K
HTML
Send an Item with 3023951 3690 2949344 2457 74607 1233
Attachment 1000K
Plain Text
Read an Item 1K 16798 29 667 9 16131 20
HTML
Read an Item 1K 7124 12 277 4 6847 8
Plain Text
Read an Item 2K 32802 43 865 12 31937 31
HTML
Read an Item 2K 13795 17 337 5 13458 12
Plain Text
Read an Item 4K 63635 74 1543 22 62092 52
HTML
Read an Item 4K 26758 31 575 8 26183 23
Plain Text
Read an Item with 31356 42 851 11 30505 31
Attachment 10K
HTML
68 Client

Read an Item with 22256 32 679 9 21577 23
Attachment 10K
Plain Text
Read an Item with 91534 100 1868 28 89666 72
Attachment 50K
HTML
Read an Item with 82300 89 1632 25 80668 64
Attachment 50K
Plain Text
Read an Item with 165042 162 2643 41 162399 121
Attachment 100K
HTML
Read an Item with 156029 155 2528 40 153501 115
Attachment 100K
Plain Text
Read an Item with 751901 640 8343 136 743558 504
Attachment 500K
HTML
Read an Item with 742678 629 8111 133 734567 496
Attachment 500K
Plain Text
Read an Item with 1485110 1242 15605 257 1469505 985
Attachment 1000K
HTML
Read an Item with 1475762 1229 15312 253 1460450 976
Attachment 1000K
Plain Text
Delete an Item 394 5 210 3 184 2
Delete1 item of 1K 231 4 126 2 105 2
item (Purge Deleted
Messages)
Delete 1 items with 231 3 126 2 105 1
1000K Attachment
(Purge Deleted
Messages)
Delete10 item of 408 4 126 2 282 2
1K item (Purge
Deleted Messages)
1000K Attachment
(Purge Deleted
Messages)
By default, Netscape Messenger stores sent items locally in the Sent folder.
Outlook Express stores sent items on a server folder. To ensure equal testing
conditions, the Netscape Sent folder was moved to the server.
Netscape Messenger does not download attachments with messages. Netscape
Messenger downloads only message bodies and then downloads the attachments
when users want access to them. The network traffic test captures at this point
included attachment download.
If a user closes the attachment and at a later point wants to access the attachment
again, Netscape Messenger downloads it again; it does not mark the attachment as
already available locally.
Graph 17 : Generic Mail Item / Netscape Messenger IMAP - Send item without
50 000
40 000
30 000
Bytes
20 000
10 000
HTM L 13 669 25 740 48 418

Plain Text 13 630 26 089 48 463
There is no gap between HTML and Plain Text using Netscape Message. It seems
to code the HTML text differently when it has no rich characters.
70 Client
Graph 18 : Generic Mail Item / Netscape Messenger IMAP - Send item with
3 500 000
3 000 000
2 500 000
2 000 000
Bytes
1 500 000
1 000 000
500 000
0 1K t ext w/ 10K 1K text w/ 50K 1K t ext w/ 100K 1K text w/ 500K 1K t ext w/ 1000K
Off ice Of fice Of fice Off ice Of fice
HTM L 43 162 165 118 315 643 1 520 309 3 024 199
Plain Text 43 219 165 115 315 698 1 520 064 3 023 951
Same analysis with or without attachment: no gap between format, when no rich
character is detected.
Graph 19 : Generic Mail Item / Netscape Messenger IMAP - Read item without
70 000
60 000
50 000
40 000
Bytes
30 000
20 000
10 000
HTM L 16 798 32 802 63 635

Plain Text 7 124 13 795 26 758
On the contrary, reading an HTML message is more bandwidth consuming than

a Plain Text message. We get the same results with Netscape Messenger or
Outlook Express.
Graph 20 : Generic Mail Item / Netscape Messenger IMAP - Read item with
1 600 000
1 400 000
1 200 000
1 000 000
Bytes
800 000
600 000
400 000
200 000
Office Office Office Office 1000K Office
HTM L 31 356 91 534 165 042 751 901 1 485 110

Plain Text 22 256 82 300 156 029 742 678 1 475 762
The difference in bandwidth consumption is mainly due to attachment size. There

is no preferred format in which to send an attachment. Attachment conversion does
not increase the difference.
6.8. Netscape Messenger POP Results

Refresh screen with 1149 16 488 8 661 8
no new message
Send an Item 1K 7684 30 6341 17 1343 13
HTML
Send an Item 1K 7621 29 6278 16 1343 13
Plain Text
Send an Item 2K 13791 39 12268 23 1523 16
HTML
Send an Item 2K 13783 39 12260 23 1523 16
Plain Text
Send an Item 4K 25181 54 23358 33 1823 21
HTML
Send an Item 4K 25231 55 23408 34 1823 21
Plain Text
72 Client

Send an Item with 22465 49 20762 30 1703 19
Attachment 10K
HTML
Send an Item with 22581 51 20818 31 1763 20
Attachment 10K
Plain Text
Send an Item with 84178 136 80735 88 3443 48
Attachment 50K
HTML
Send an Item with 84173 136 80731 88 3442 48
Attachment 50K
Plain Text
Send an Item with 160458 245 154735 159 5723 86
Attachment 100K
HTML
Send an Item with 160332 243 154730 159 5602 84
Attachment 100K
Plain Text
Send an Item with 768848 1086 746385 721 22463 365
Attachment 500K
HTML
Send an Item with 768784 1085 746321 720 22463 365
Attachment 500K
Plain Text
Send an Item with 1528689 2138 1485166 1422 43523 716
Attachment 1000K
HTML
Send an Item with 1528628 2137 1485164 1422 43464 715
Attachment 1000K
Plain Text
Read an Item 1K 17220 37 922 15 16298 22
HTML
Read an Item 1K 8305 29 802 13 7503 16
Plain Text
Read an Item 2K 33133 50 1102 18 32031 32
HTML
Read an Item 2K 15037 35 922 15 14115 20
Plain Text
Read an Item 4K 63070 74 1342 22 61728 52
HTML
Read an Item 4K 27571 44 982 16 26589 28
Plain Text
Read an Item with 31522 48 982 16 30540 32
Attachment 10K
HTML
Read an Item with 22657 40 922 15 21735 25
Attachment 10K
Plain Text

Read an Item with 90980 97 1642 27 89338 70
Attachment 50K
HTML
Read an Item with 82114 89 1522 25 80592 64
Attachment 50K
Plain Text
Read an Item with 164481 159 2482 41 161999 118
Attachment 100K
HTML
Read an Item with 155495 149 2242 37 153253 112
Attachment 100K
Plain Text
Read an Item with 751361 637 8122 135 743239 502
Attachment 500K
HTML
Read an Item with 742495 629 8002 133 734493 496
Attachment 500K
Plain Text
Read an Item with 1484362 1236 15322 255 1469040 981
Attachment 1000K
HTML
Read an Item with 1475556 1229 15262 254 1460294 975
Attachment 1000K
Plain Text
Graph 21 : Generic Mail Item / Netscape Messenger POP - Send item without
30 000
25 000
20 000
Bytes
15 000
10 000
5 000
HTM L 7 684 13 791 25 181

Plain Text 7 621 13 783 25 231
74 Client
For Netscape in IMAP mode, there is no difference in sending in HTML or Plain

Text.
Graph 22 : Generic Mail Item / Netscape Messenger POP - Send item with
attachment(comparative: HTML, Plain Text)
1 600 000
1 400 000
1 200 000
Bytes 1 000 000

800 000
600 000
400 000
200 000
0 1K t ext w/ 10K 1K t ext w/ 50K 1K t ext w/ 100K 1K text w/ 500K 1K text w/
Off ice Off ice Off ice Off ice 1000K Off ice
HTM L 22 465 84 178 160 458 768 848 1 528 689

Plain Text 22 581 84 173 160 332 768 784 1 528 628
There were similar results with or without attachments. It made no difference in

traffic consumption for different formats when no rich characters were detected.
Graph 23 : Generic Mail Item / Netscape Messenger POP - Read item without
70 000
60 000
50 000
40 000
Bytes
30 000
20 000
10 000
HTM L 17 220 33 133 63 070

Plain Text 8 305 15 037 27 571
HTML messages consume more bandwidth than Plain Text messages, for both
Netscape Messenger and Outlook Express.
1 600 000
1 400 000
1 200 000
1 000 000
Bytes
800 000
600 000
400 000
200 000
HTM L 31 522 90 980 164 481 751 361 1 484 362

Plain Text 22 657 82 114 155 495 742 495 1 475 556
The difference in bandwith used to send these message types is mainly due to
attachment size. There is no preferred format in which to send attachments.
Attachment conversion does not significantly increase the difference.
For all the below tests, we use the default view which is “Messages” in Inbox.
6.9.1. Tests Details

MSG1: New Message, first time in the windows profile
1. Open the windows session
2. Run Outlook Web Access (http://LONDON-04/exchange)
3. In Inbox, click New in Outlook toolbar.
MSG2: Attachments
1. In a blank new message, click Attachments.
REF0, 1, 2, 3: Check for new messages when 0, 1, 5 or 10 new message(s)

1. Have x new message(s) where x is 0 (1, 5, 10)
2. Click Check for new messages.
76 Client
SH1, 2, 3: Send a message with 1 KB, 2 KB, and 4 KB of text

1. Open the x KB reference .txt file where x is 1 (2, 4).
2. Copy the text.
4. In the To field, enter a recipient other than yourself.
5. In the Subject field, enter Hx KB0000 where x is 1 (2, 4).
6. Paste the text in the main message body.
7. Click Send.
SH4, 5, 6, 7, 8: Send a message with 1 KB of text and an attachment of 10, 50, 100,
500, and 1000 KB.
1. Open the x KB reference .txt file where x is 1 (2, 4).
2. Copy the text.
4. In the To field, enter a recipient other than yourself.
5. In the Subject field, enter H1 KBx where x is 0010 (0050, 0100, 0500,
1000).
6. Paste the text in the main message body.
8. Click Send.
RH1, 2, 3: Open a message with 1 KB, 2 KB, and 4 KB of text

1. Open the message, Hx KB0000, where x is 1 (2, 4).
RH4: Open a message with 1 KB of text and an attachment of any size

1. Open the message, H1 KB0010 without opening the attachment.
OH4, 5, 6, 7, 8: Open the 10 KB, 50 KB, 100 KB, 500 KB, and 1000 KB attachment in
the 1 KB message.
1. Open the message, H1 KB0010, without opening the attachment.
D1: Delete a message (with no attachment)

1. Select a message among many messages in the current folder with no
attachment.
2. Press Delete.
D2: Delete a message (with one attachment)

1. Select a message among many messages in the current folder with one
attachment.
2. Press Delete.
D3: Delete the last message (with no attachment)

1. Select the last message in the current folder with no attachment.
2. Press Delete.
D4: Delete the last message (with one attachment)

1. Select the last message in the current folder with one attachment.
2. Press Delete.
PRR1: Open a read receipt.

1. Open a read receipt.
PDR1: Open a delivery receipt

1. Open a delivery receipt.
DI1: Empty Deleted Items containing one 1 KB message

1. Put one 1 KB message in Deleted Items folder.
2. Empty Deleted Items folder.
DI2: Empty Deleted Items containing one 1 KB message with one 1000 KB attachment
1. Put one 1 KB message containing one 1000 KB attachment in Deleted
Items folder.
2. Empty Deleted Items folder.
DI3: Empty Deleted Items containing ten 1 KB messages

1. Put ten 1 KB messages in Deleted Items folder.
2. Click Empty Deleted Items Folder.
DI4: Empty Deleted Items containing ten 1 KB messages with one 1000 KB
attachment each
1. Put ten 1 KB messages each containing a 1000 KB attachment in Deleted
Items folder.
2. Click Empty Deleted Items Folder.
FOL1: Create Folder (only show dialog)

1. Click Folder in the New tool
FOL2: Create Folder (full)

1. Click Folder in the New tool then enter Folder1 then OK
78 Client

Open New 159214 196 11745 77 147469 119
Message display
Open Attach display 5844 11 939 5 4905 6
no new message
1 message
5 messages
Refresh Screen with 9401 12 3254 6 6147 6
10 messages
Send an Item 1K 10463 14 9168 9 1295 5
Send an Item 2K 18167 24 16630 16 1537 8
Send an Item 4K 32181 36 30404 24 1777 12
Send an Item with 50017 69 35135 41 14882 28
Attachment 10K
Send an Item with 93770 110 78167 40 15603 70
Attachment 50K
Send an Item with 147397 154 131251 105 16146 49
Attachment 100K
Send an Item with 579617 561 557051 405 22566 156
Attachment 500K
Send an Item with 1119764 1076 1088794 780 30970 296
Attachment 1000K
Read an Item 1K 16183 18 627 6 15556 12
Read an Item 2K 23559 22 627 6 22932 16
Read an Item 4K 37819 36 843 10 36976 26
Read an Item with 16615 18 627 6 15988 12
Attachment
Open the 10K 11014 11 588 4 10426 7
Attachment
Open the 50K 54744 53 1290 17 53454 36
Attachment
Open the 100K 108867 107 2321 36 106546 71
Attachment
Open the 500K 540387 513 9067 161 531320 352
Attachment
Open the 1000K 1079429 1023 17708 321 1061721 702
Attachment
Delete an Item (last 7173 14 5606 8 1567 6
item in the view)
Delete an item (10 16661 23 8079 12 8582 11
items remaining)

Delete an Item with 7173 14 5606 8 1567 6
Attachment (last)
Delete an Item with 16661 23 8079 12 8582 11
Attachment (10
items remaining)
Open a read receipt 7015 9 519 3 6496 6
Open a delivery 7172 10 578 4 6594 6
receipt
Delete1 item of 1K 834 5 508 3 326 2
item
1000K Attachment
Delete10 item of 1K 834 5 508 3 325 2
item
1000K Attachment
FOL1: Show Create 19980 28 5706 13 14274 15
New folder
FOL2: Create a 30736 45 10793 22 19943 23
folder
Showing the new message form generates traffic only the first time in the windows
profile. For any further new messages, no more traffic occurs. The network traffic
that occurs when you delete items in a folder depends on the number of items that
remain in the folder. When you delete items, the whole folder content is refreshed.
The fewer messages that remain, the less traffic is generated. But, more traffic is
generated when you delete the last item than when some items stay. Deleting does
not depend no item size.
Attachments are upload as soon as you click Attach in the Attachments dialog.
Clicking Attachments check names which generates more traffic if you already
filled a large text in the body. Attachments are download on demand.
It is more efficient to refresh your view with Check for new messages than with
Internet Explorer Refresh command (F5).
80 Client
6.10. Mail Item Analysis
HTML is the more–consuming bandwidth format, regardless of the client.

Forwarding messages does not always download attachments from the original
message. It depends on the format of the original message. If the message format is
Rich Text, the attachment is downloaded when the message forwards. If the format
is HTML or Plain Text, the attachment is never downloaded; it stays on the server.
For instance, if you receive a plain text message with a 1–MB document, and you
forward this mail, you will not download the 1–MB document from the server. This
functionality is specific to Outlook 2000 because it handles messages as RTF
format, and downloads attachments when you forward a message with attachments.
In POP3 mode, attachments fully download as soon as you open or read the
message. This characteristic is essential for POP3 protocol. So, forwarding is
similar to sending a new message.
In IMAP mode, attachment management depends on the client. Outlook
Express 5.x always downloads the whole message, including attachments.
Netscape Messenger does not. In tests, Netscape Messenger fully downloaded
messages with 10–KB attachments when the message was forwarded, whereas it
only downloads 50–KB (and greater) attachments when you want to open the
attachment. Large attachments download immediately when you click Forward.
There is no global difference in the way Outlook 2000 or Outlook 97 handles RTF
messages. Nor is there any essential difference between the two MAPI clients.
Graph 25 : Generic Mail Item - Send item in RTF format without attachment
(comparative: Outlook 2000, Outlook 97)
9 000
8 000
7 000
6 000
Bytes
5 000
4 000
3 000
2 000
1 000
0 1K wit h no att achment 2K wit h no att achment 4K wit h no att achment
Out look 2000 5 354 6 442 8 386

Out look 97 5 358 6 394 8 418
In practice, there is no difference between the two MAPI clients.

Graph 26 : Generic Mail Item - Send item in RTF format with attachment
1 200 000
1 000 000
800 000
Bytes
600 000
400 000
200 000
0 1K w/ 50K 1K w/ 100K 1K w/ 500K 1K w/ 1000K

1K w/ 10K Of fice
Of fice Off ice Of fice Off ice
att achment
at t achment at tachment at t achment at tachment
Out look 2000 21 192 65 272 120 792 564 422 1 118 554
Out look 97 35 106 85 004 120 352 564 102 1 118 234
In practice, there is – again - no difference between the two MAPI clients, even
if you send attachments or not.
Graph 27 : Generic Mail Item - Send item in HTML format without attachment
(comparative: Outlook 2000, Outlook Express IMAP / POP, Netscape IMAP / POP,
OWA)
140 000
120 000
100 000
80 000
Bytes
60 000
40 000
20 000
0 HTM L 1K message HTM L 2K message HTM L 4K message
Outlook 2000 15 794 27 822 50 594

Outlook Express IM AP 34 344 67 005 127 707
Netscape IM AP 13 669 25 740 48 418
Outlook Express POP 17 149 34 403 64 814
Netscape POP 7 684 13 791 25 181
OWA 10463 18167 32181
Outlook Express is the most bandwidth consuming among all the tested clients
when messages are sent in HTML format.
IMAP clients consume more bandwidth than any other client—about twice the
traffic consumed by POP3 clients. Outlook Web Access generates the least network
traffic for HTML-formatted messages. Therefore, Outlook Web Access is the best
client for reading large text message.
82 Client
Graph 28 : Generic Mail Item - Send item in HTML format with attachment
OWA)
3 500 000
3 000 000
2 500 000
Bytes
2 000 000
1 500 000
1 000 000
500 000
0 1K text w/ 1K text w/ 1K text w/ 1K text w/ 1K text w/
10K 50K 100K 500K 1000K
Outlook 2000 31 104 75 244 130 644 574 394 1 128 466
Outlook Express IMAP 63 793 184 331 332 972 1 523 040 3 009 231
Netscape IMAP 43 162 165 118 315 643 1 520 309 3 024 199
Outlook Express POP 32 857 93 126 166 506 762 390 1 505 455
Netscape POP 22 465 84 178 160 458 768 848 1 528 689
OWA 50 017 93 770 147 397 579 617 1 119 764
The difference between Outlook Express and the other clients is not so visible
because attachments create most of the traffic. The difference between IMAP and
POP clients is still there. MAPI clients and Outlook Web Access are very close;
they are the lowest consuming clients for HTML messages.
Graph 29 : Generic Mail Item - Send item in Plain Text format without attachment
(comparative : Outlook 2000, OExpress IMAP / POP, Netscape IMAP / POP)
60 000
50 000
40 000
Bytes
30 000
20 000
10 000
Outlook 2000 5 958 6 950 9 018

Netscape IM AP 13 630 26 089 48 463
Netscape POP 7 621 13 783 25 231
There is no bad client. The gap between IMAP and POP clients is clearly visible
(a two ratio). The MAPI client is the best one.
Graph 30 : Generic Mail Item - Send item in Plain Text format with attachment
(comparative: Outlook 2000, Outlook Express IMAP / POP, Netscape IMAP / POP)
3 500 000
3 000 000
2 500 000
2 000 000
Bytes
1 500 000
1 000 000
500 000
0 1K text w/ 1K text w/ 1K text w/ 1K t ext w/ 1K t ext w/
10K 50K 100K 500K 1000K
Outlook 2000 21 120 65 316 120 748 564 498 1 118 630
Outlook Express IM AP 45 710 166 267 315 138 1 505 086 2 991 219
Netscape IM AP 43 219 165 115 315 698 1 520 064 3 023 951
Netscape POP 22 581 84 173 160 332 768 784 1 528 628
The difference in network traffic between IMAP and POP clients is clearly visible.
The MAPI client is the lowest consumer of bandwidth.
Graph 31 : Generic Mail Item - Read item in RTF format without attachment
9 000
8 000
7 000
6 000
Bytes
5 000
4 000
3 000
2 000
1 000
Out look 2000 5 084 6 060 8 346

Out look 97 5 440 6 508 8 734
There is no difference between Outlook clients even for read or send messages.
84 Client
Graph 32 : Generic Mail Item - Read and Open Item in RTF format with attachment
1 200 000
1 000 000
800 000
Bytes
600 000
400 000
200 000
0 1K t ext w/ 10K 1K t ext w/ 50K 1K t ext w/ 100K 1K t ext w/ 500K 1K text w/ 1000K
Outlook 2000 21 948 66 890 122 308 566 608 1 121 334
Outlook 97 25 406 68 334 124 472 574 772 1 137 058
Same analysis with or without attachment.
Graph 33 : Generic Mail Item - Read item in HTML format without attachment
OWA)
80 000
60 000
Bytes
40 000
20 000

Outlook 2000 17 212 28 924 68 714
Outlook Express IMAP 16 800 32 711 62 882
Netscape IMAP 16 798 32 802 63 635
Netscape POP 17 220 33 133 63 070
OWA 16183 23559 37819
Reading a HTML message is equal whatever the client, except for OWA, which
take more traffic for small message. Surprisingly, OWA is the best client for
reading large text message.
Graph 34 : Generic Mail Item - Read and Open item in HTML format with attachment
OWA)
1 500 000
1 000 000
Bytes
500 000

10K 50K 100K 500K 1000K
Outlook 2000 33 940 78 882 134 300 578 600 1 133 326
Outlook Express IMAP 31 220 90 973 164 887 755 728 1 493 539
Netscape IMAP 31 356 91 534 165 042 751 901 1 485 110
Netscape POP 31 522 90 980 164 481 751 361 1 484 362
OWA 38498 82228 136351 567871 1 106 913
Reading a HTML message with attachment is equal for IMAP and POP clients.
OWA and MAPI clients are the best. We find the same evolution for OWA than
in the previous chart: OWA is the best client for reading large messages.
Graph 35 : Generic Mail Item - Read item in Plain Text format without attachment
(comparative: Outlook 2000, Outlook Express IMAP / POP, Netscape IMAP / POP)
30 000
25 000
20 000
Bytes
15 000
10 000
5 000
Outlook 2000 5 116 6 092 9 974

Netscape IM AP 7 124 13 795 26 758
Netscape POP 8 305 15 037 27 571
There is no gap between clients when reading Plain Text messages.

86 Client
Graph 36 : Generic Mail Item - Read and Open item in Plain Text format with
attachment (comparative: Outlook 2000, Outlook Express IMAP / POP, Netscape
IMAP / POP)
1 500 000
1 000 000
Bytes
500 000

10K 50K 100K 500K 1000K
Out look 2000 21 980 66 922 122 340 566 640 1 121 366
Out look Express IM AP 22 355 82 107 155 902 746 802 1 484 499
Net scape IM AP 22 256 82 300 156 029 742 678 1 475 762
Out look Express POP 23 096 82 794 156 715 745 275 1 485 356
Net scape POP 22 657 82 114 155 495 742 495 1 475 556
When reading Plain Text messages with attachment, there is no gap between
clients, except for the MAPI client, which is the most bandwidth safe.
7. Calendaring, Contacts and

Task items Tests
These tests provide traffic measurements generated by Calendar, Contacts, and

Tasks items. These tests were performed on the following clients:
• Microsoft Outlook Web Access (no Tasks items)
The following list describes the general tasks that each mail client performed for
the tests
• Open calendar Initial view of the calendar, initiated by clicking the Calendar
icon.
• Open calendar appointment Initial view of a new appointment, initiated by
clicking New in the Calendar view.
• Add a calendar item Create a four-hour meeting in the Calendar (from 8 a.m. to
12 p.m. any day) using “Meeting” as the description.
• Modify a calendar item Move the above meeting to a new time slot (12 p.m. to
4 p.m.) on the same day; rename the meeting “Another Meeting”.
• Delete a calendar item Delete the above appointment.
• Open contacts Initial view of contacts, initiated by clicking the Contacts item.
• Open contact form Initial view of the Contact form, initiated by clicking New in
the Contact view.
• Add a contact item Create a contact with full name, company name, two
telephone numbers, business address, and an e-mail address. Click Save and
Close.
• Modify a contact item Rename the contact.
• Delete a contact item Delete the contact created above.
• Open tasks Initial view of tasks, initiated by clicking the Tasks item.
88 Client
• Open task form Initial view of the task form, initiated by clicking New in the
Task view.
• Add a task item Create a task called “Task”.
• Modify a task item Rename it to “Another task”. Click Save and Close.
• Delete a task item Delete the above task.
7.2. Test Details

CAL1: Open Calendar folder
1. Start Outlook.
2. Click Calendar in Outlook bar.
CAL2: New appointment: first Outlook session

1. Start Outlook.
3. Create a new appointment.
CAL3: Add an appointment

2. Create a new appointment.
3. Fill required boxes, and then click Save and Close.
CAL4: Modify an appointment

2. Open an appointment, modify the subject, and then click Save and Close.
CAL5: Delete an appointment

2. Select an appointment.
3. Delete the appointment.
CTC1: Open Contacts folder

1. Start Outlook.
2. Click Contacts in Outlook bar.
CTC2: New contact: first Outlook session

1. Start Outlook.
3. Create a new contact.
CTC3: Add a contact

2. Create a new contact.
CTC4: Modify a contact

2. Open a contact, and then modify the company.
3. Click Save and Close.
CTC5: Delete a contact

2. Select a contact.
3. Delete the contact.
TSK1: Open Tasks folder

1. Start Outlook.
2. Click Tasks in Outlook bar.
TSK2: New Task: first Outlook session

1. Start Outlook.
3. Create a new Task.
TSK3: Add a Task

2. Create a new Task.
TSK4: Modify a Task

2. Open a Task, modify the subject, and then click Save and Close.
TSK5: Delete a Task

2. Select a task
3. Delete the task.
90 Client
Received by
Client

Open calendar folder 4518 14 1554 8 2964 6
Open appointment form 1st
4018 8 3258 5 760 3
time in session
Add an appointment 7498 20 5150 11 2348 9
Modify an appointment 6278 15 2444 8 3834 7
Delete an appointment 6938 20 2676 11 4262 9
Open contacts folder 1124 5 648 3 476 2
Open contact form 1st time
3922 8 3162 5 760 3
in session
Add a contact 3228 9 1828 5 1400 4
Modify a contact 5474 9 2090 5 3384 4
Delete a contact 1964 9 724 5 1240 4
Open Tasks folder 1092 5 616 3 476 2
Open task form 1st time in
3534 10 2672 6 862 4
session
Add a task 3876 13 1840 7 2036 6
Modify a task 4922 13 1846 7 3076 6
Delete a task 1632 7 558 4 1074 3
Received by
Client

Open appointment form 1st
3452 8 2746 5 706 3
time in session
Open contact form 1st time
632 3 354 2 278 1
in session
Add a contact 4824 17 2894 10 1930 7
Received by
Client

Modify a contact 5422 11 1968 6 3454 5
Delete a contact 1284 5 392 3 892 2
Open tasks folder 584 3 354 2 230 1
Open task form 1st time in
568 3 306 2 262 1
session
Add a task 4004 13 1952 7 2052 6
Modify a task 4414 11 1600 6 2814 5
Delete a task 1440 8 462 5 978 3
7.5. OWA Results
Received by
Client

Open appointment form 151578 168 8279 61 143299 107
Open contact form 60540 84 6997 36 53543 48
Add a contact 7120 13 4729 8 2391 5
Modify a contact 32102 37 5850 16 26252 21
Delete a contact 12009 21 8666 12 3343 9
92 Client
7.6. Calendaring, Contacts, and Tasks Analysis
Graph 37 : Calendar actions (comparative : Outlook 2000, Outlook 97, OWA)
40000
35000
30000
25000
Bytes
20000
15000
10000
5000
0 Open calendar Modify a calendar Delete a calendar
Add a calendar item
folder item item
Outlook 2000 4518 7498 6278 6938
Outlook 97 1552 7472 5798 5830
OWA 35888 15296 39127 12445
Calendar actions are very similar with both MAPI clients, except when opening the
calendar folder where Outlook 2000 has more features than Outlook 97. Outlook
Web Access is more bandwidth–consuming. This delta is visible when displaying
folder or opening item. Outlook Web Access downloads blank new form (message,
appointment, contact) only once in the windows profile; then it caches it in the
Internet Explorer cache folder. Further requests do not create additional traffic. It is
very similar to MAPI clients which get the form design from locally-stored
templates.
Graph 38 : Contact actions (comparative : Outlook 2000, Outlook 97, OWA)
35000
30000
25000
20000
Bytes
15000
10000
5000
0 Open calendar Modify a calendar Delete a calendar
Add a calendar item
folder item item
Outlook 2000 1124 3228 5474 1964
Outlook 97 632 4824 5422 1284
OWA 21898 7120 32102 12009
Similar conclusions can be drawn with regard to contact functionality. Outlook

Web Access downloads more data than MAPI clients to display the contacts forms.
Graph 39 : Task actions (comparative : Outlook 2000, Outlook 97)
5000
4000
3000
Bytes
2000
1000
0 M odify a calendar Delete a calendar

Open calendar folder Add a calendar item
item it em
Out look 2000 1092 3876 4922 1632

Out look 97 584 4004 4414 4978
Outlook 2000 and Outlook 97 are equivalent in tasks operations and related
network consumption.
94 Client
8. Public Folders
The public folder tests provide traffic measurements generated by public folders.
They include MAPI and NNTP access. The tests were performed on the following
clients.
• Microsoft Outlook Express 5.0 (with an NNTP account)
• Netscape Messenger 4.7 (with an NNTP account)
• Microsoft Outlook Web Access
For MAPI-client testing, mail was not sent to a public folder’ instead, the New Post
form was used.
The following list describes the general tasks that each mail client performed for
the tests.
• Connection When the user clicks Public Folders, an initial connection occurs
with the associated public folders server.
• Hierarchy listing When the user double-clicks All Public Folders, the hierarchy
of public folders, within the given organization appears.
• Post an item x KB (1 KB, 2 KB, 4 KB) Post a message containing x KB of text.
Include a simple subject indicating the contents of the message, such as “x KB of
text.” Create posts using the following formats.
• RTF
• HTML
• Plain Text
• Post an item with attachment of x KB (10 KB, 50 KB, 100 KB, 500 KB,
1,000 KB) Post a message containing an x KB attachment with no other text.
Include a simple subject indicating the contents of the message, such as “x KB
Attachment”. Create posts using the following formats.
• RTF
• HTML
• Plain Text
• Read an item: x KB (with or without attachment) Open the items posted in the
previous two tests.
• Open the attachment within items Open the attachments posted in the earlier
test.
• Delete an item with or without attachment Delete the items posted in the
earlier test.
8.2. Test Details
The test steps are identical to the tests for generic mail items.
In addition, the following tests were also run.
CNX1: Connection, first Outlook session
1. Start Outlook.
2. Show Folder view.
3. Double-click Public Folders.
CNX2: Connection, second Outlook session

PF1, PF5, PF10: Open All Public Folders, with 1, 5, or 10 public folder(s), first Outlook
session
1. Start Outlook.
4. Double-click All Public Folders containing 1, 5, or 10 public folder(s).
PF2, PF6, PF11: Open All Public Folders, with 1, 5, or 10 public folders, second
Outlook session
1. Start Outlook.
4. Double-click All Public Folders containing 1, 5, or 10 public folder(s).
The following tests were only run on NNTP clients.

96 Client
PF1, PF5, PF10: Refresh Newsgroups List, with 1, 5, or 10 public folders, first client
session
1. Start the NNTP client.
2. Right-click the newsgroups server, and then choose Newsgroups.
3. Refresh the newsgroups list (In Outlook Express, click Reset List).
PF2, PF6, PF11: Open All Public Folders, with 1, 5, or 10 public folders, second client
session
This test was run after the PF1, PF5, and PF10 test.
1. Right-click the newsgroups server, and then choose Newsgroups.
2. Refresh the newsgroups list (In Outlook Express, click Reset List).
8.3. Outlook 2000 results

CNX1: 5728 31 2806 16 2922 15
Connection (1st
time in session)
CNX2: 2260 13 1200 7 1060 6
Connection (2nd
time)
PF1: Open “All 1152 7 590 4 562 3
Public Folders”
with 1 public
folder (1st time)
PF2: Open “All 488 3 258 2 230 1
Public Folders”
with 1 public
folder (2nd time)
PF5: Open “All 1264 7 590 4 674 3
Public Folders”
with 5 public
folder (1st time)
PF6: Open “All 584 3 258 2 326 1
Public Folders”
with 5 public
folder (2nd time)

PF10: Open “All 1392 7 590 4 802 3
Public Folders”
with 10 public
folder (1st time)
PF11: Open “All 712 3 258 2 454 1
Public Folders”
with 10 public
folder (2nd time)
Send an Item 1K 2624 7 1870 4 754 3
RTF
Send an Item 1K 3744 7 2990 4 754 3
HTML
Send an Item 1K 2656 7 1902 4 754 3
Plain Text
Send an Item 2K 2768 7 2014 4 754 3
RTF
Send an Item 2K 5094 8 4340 5 754 3
HTML
Send an Item 2K 2800 7 2046 4 754 3
Plain Text
Send an Item 4K 3056 7 2302 4 754 3
RTF
Send an Item 4K 7982 12 7104 7 878 5
HTML
Send an Item 4K 3088 7 2334 4 754 3
Plain Text
Send an Item 17388 24 16134 15 1254 9
with Attachment
10K RTF
Send an Item 18416 23 17162 14 1254 9
with Attachment
10K HTML
Send an Item 17344 23 16090 14 1254 9
with Attachment
10K Plain Text
Send an Item 62130 72 59734 46 2396 26
with Attachment
50K RTF
Send an Item 63158 71 60762 45 2396 26
with Attachment
50K HTML
Send an Item 62086 71 59690 45 2396 26
with Attachment
50K Plain Text
98 Client

Send an Item 117422 130 113556 83 3866 47
with Attachment
100K RTF
Send an Item 140556 152 136130 97 4426 55
with Attachment
100K HTML
Send an Item 117432 130 113512 82 3920 48
with Attachment
100K Plain Text
Send an Item 560272 605 544256 383 18116 222
with Attachment
500K RTF
Send an Item 561300 604 545184 382 16116 222
with Attachment
500K HTML
Send an Item 580492 608 544112 382 16380 226
with Attachment
500K Plain Text
Send an Item 1113276 1199 1081862 758 31414 441
with Attachment
1000K RTF
Send an Item 1114364 1199 1082950 758 31414 441
with Attachment
1000K HTML
Send an Item 1113172 1197 1081758 756 31414 441
with Attachment
1000K Plain
Text
Open an Item 1K 3486 4 898 2 2588 2
RTF
Open an Item 1K 5084 8 1156 4 3928 4
HTML
Open an Item 1K 3502 4 898 2 2604 2
Plain Text
Open an Item 2K 3630 4 898 2 2732 2
RTF
Open an Item 2K 6450 9 1172 4 5278 5
HTML
Open an Item 2K 3662 4 898 2 2764 2
Plain Text
Open an Item 4K 3902 4 898 2 3004 2
RTF
Open an Item 4K 10762 13 1962 5 8800 8
HTML

Open an Item 4K 3988 5 898 2 3090 3
Plain Text
Open an Item 1K 20366 32 2530 14 17836 18
with Attachment
10K RTF
Open an Item 1K 21920 35 2744 15 19176 20
with Attachment
10K HTML
Open an Item 1K 20398 32 2546 14 17852 18
with Attachment
10K Plain Text
Open an Item 1K 65324 72 3596 24 61728 48
with Attachment
50K RTF
Open an Item 1K 66862 75 3794 25 63068 50
with Attachment
50K HTML
Open an Item 1K 65340 72 3596 24 61744 48
with Attachment
50K Plain Text
Open an Item 1K 120742 121 4856 36 115886 85
with Attachment
100K HTML
Open an Item 1K 122280 124 5054 37 117226 87
with Attachment
100K HTML
Open an Item 1K 120758 121 4856 36 115902 85
with Attachment
100K Plain Text
Open an Item 1K 565010 521 15340 136 549670 385
with Attachment
500K RTF
Open an Item 1K 566580 524 15554 137 551026 387
with Attachment
500K HTML
Open an Item 1K 565010 521 15356 136 549702 385
with Attachment
500K Plain Text
Open an Item 1K 1119768 1020 28376 260 1091392 760
with Attachment
1000K RTF
Open an Item 1K 1121306 1023 28574 261 1092732 762
with Attachment
1000K HTML
100

Open an Item 1K 1119784 1020 28376 260 1091408 760
with Attachment
1000K Plain
Text
Delete an Item 900 5 392 3 508 2
Delete an Item 916 5 392 3 524 2
with Attachment
After the initial connection from the Outlook client to the Public Folder server,
each subsequent connection (in the same MAPI session) generated traffic. Clicking
Public Folders does not generate traffic based on the total number of public
folders. Instead, the traffic generated is dependent on the number of folders within
the Favorites folder. The status of the folders in the Favorites folder updates to
reflect the number of unread messages.
Clicking All Public Folders generates traffic based on the number of first-level
folders; sub folders do not generate traffic at this point. The first connection to All
Public Folders generates more traffic than subsequent connections, and these only
check for changes within the hierarchy.
For generic mail items, deleting a publication item is not effected by the message
size; rather, traffic levels rely on the existence of attachments. The traffic is equal
to one or many attachments.
Graph 40 : Public Folder / Outlook 2000 - Send Item without attachment
8 000
7 000
6 000
5 000
Bytes
4 000
3 000
2 000
1 000
RTF 2 624 2 768 3 056

HTM L 3 744 5 094 7 982
Plain Text 2 656 2 800 3 088
HTML posts are larger than any other formats. In HTML tests, the text used was
not the same as in the generic mail item tests; instead, text from Readme files were
used. HTML text conversion depends on the form of the text format: the more line
breaks, the bigger the converted text.
Graph 41 : Public Folder / Outlook 2000 - Send Item with attachment (comparative:
RTF, HTML, Plain Text)
1 200 000
1 000 000
800 000
Bytes
600 000
400 000
200 000
0 1K text w/ 10K 1K text w/ 50K 1K t ext w/ 1K text w/ 1K text w/

Off ice Off ice 100K Of fice 500K Of fice 1000K Off ice
att achment att achment att achment at tachment at tachment
RTF 17 388 62 130 117 422 560 272 1 113 276

HTM L 18 416 63 158 140 556 561 300 1 114 364
Plain Text 17 344 62 086 117 432 560 492 1 113 172
Message text formats do not impact network traffic when attachments are inserted.
The small difference with 100 KB in HTML can be due to an upload, which begins
when a message posts.
Graph 42 : Public Folder / Outlook 2000 - Read item without attachment

12 000
10 000
8 000
Bytes
6 000
4 000
2 000
RTF 3 486 3 630 3 902

HTM L 5 084 6 450 10 762
Plain Text 3 502 3 662 3 988
Analysis on reading publications is similar to the sending publications analysis:

HTML format is bigger.
102
Graph 43 : Public Folder / Outlook 2000 - Read item with attachment (comparative:
RTF, HTML, Plain Text)
1 200 000
1 000 000
800 000
Bytes
600 000
400 000
200 000
0 1K text w/ 10K 1K text w/ 50K 1K text w/ 1K text w/ 1K text w/

Office Office 100K Office 500K Office 1000K Office
attachment attachment attachment at tachment attachment
RTF 20 366 65 324 120 742 565 010 1 119 768

HTM L 21 920 66 862 122 280 566 580 1 121 306
Plain Text 20 398 65 340 120 758 565 058 1 119 784
Analysis on reading publications with attachment is similar to the sending

publications analysis: HTML format is bigger.
8.4. Outlook 97 results

CNX1: Connection 5616 31 2774 15 2842 16
(1st time in session)
CNX2: Connection 2212 13 1168 7 1044 6
(2nd time)
PF1: Open “All 804 5 424 3 380 2
Public Folders”
with 1 public
folder (1st time)
PF2: Open “All 488 3 258 2 230 1
Public Folders”
with 1 public
folder (2nd time)

PF5: Open “All 916 5 424 3 492 2
Public Folders”
with 5 public
folder (1st time)
PF6: Open “All 584 3 258 2 326 1
Public Folders”
with 5 public
folder (2nd time)
PF10: Open “All 1028 5 424 3 604 2
Public Folders”
with 10 public
folder (1st time)
PF11: Open “All 712 3 258 2 454 1
Public Folders”
with 10 public
folder (2nd time)
Send an Item 1K 2624 7 1870 4 754 3
RTF
Send an Item 2K 2768 7 2014 4 754 3
RTF
Send an Item 4K 3040 7 2286 4 754 3
RTF
Send an Item with 17312 23 16058 14 1254 9
Attachment 10K
RTF
Send an Item with 62054 71 59658 45 2396 26
Attachment 50K
RTF
Send an Item with 117400 130 113480 82 3920 48
Attachment 100K
RTF
Send an Item with 560574 611 544080 382 16494 229
Attachment 500K
RTF
Send an Item with 1113794 1209 1081786 757 32008 452
Attachment 1000K
RTF
Open an Item 1K 3214 4 786 2 2428 2
RTF
Open an Item 2K 3358 4 786 2 2572 2
RTF
Open an Item 4K 3646 4 786 2 2860 2
RTF
104

Open an Item 1K 20610 35 2582 17 18028 18
with Attachment
10K RTF
Open an Item 1K 66168 85 4248 37 61920 48
with Attachment
50K RTF
Open an Item 1K 122306 146 6228 61 116078 85
with Attachment
100K RTF
Open an Item 1K 572606 646 22728 261 549878 385
with Attachment
500K RTF
Open an Item 1K 1134988 1273 43308 511 1091680 762
with Attachment
1000K RTF
Delete an Item 900 5 392 3 508 2
Delete an Item 916 5 392 3 524 2
with Attachment
8.5. Outlook Express NNTP Results

PF1: Open “All 713 9 312 5 401 4
Public Folders” with
1 public folder (1st
time)
PF2: Open “All 224 3 123 2 101 1
1 public folder (2nd
time)
PF5: Open “All 841 9 312 5 529 4
time)

PF6: Open “All 352 3 123 2 229 1
time)
PF10: Open “All 987 9 312 5 675 4
time)
PF11: Open “All 498 3 123 2 375 1
time)
Send an Item 1K 5490 21 4704 11 786 10
HTML
Send an Item 1K 2831 18 2099 9 732 9
Plain Text
Send an Item 2K 8383 22 7597 12 786 10
HTML
Send an Item 2K 3906 19 3174 10 732 9
Plain Text
Send an Item 4K 14332 31 13330 17 1002 14
HTML
Send an Item 4K 6169 23 5329 12 840 11
Plain Text
Send an Item with 20002 36 18946 21 1056 15
Attachment 10K
HTML
Send an Item with 17297 33 16295 19 1002 14
Attachment 10K
Plain Text
Send an Item with 80363 103 77957 63 2406 40
Attachment 50K
HTML
Send an Item with 79028 101 76622 61 2406 40
Attachment 50K
Plain Text
Send an Item with 154802 184 150776 114 4026 70
Attachment 100K
HTML
Send an Item with 155249 185 151169 114 4080 71
Attachment 100K
Plain Text
106

Send an Item with 750406 834 733474 525 16932 309
Attachment 500K
HTML
Send an Item with 764598 850 747288 534 17310 316
Attachment 500K
Plain Text
Send an Item with 1494197 1646 1461011 1036 33186 610
Attachment 1000K
HTML
Send an Item with 1527386 1724 1491230 1059 36156 665
Attachment 1000K
Plain Text
Open an Item 1K 4415 6 186 3 4229 3
HTML
Open an Item 1K 1801 4 126 2 1675 2
Plain Text
Open an Item 2K 7423 9 246 4 7177 5
HTML
Open an Item 2K 2822 4 125 2 2697 2
Plain Text
Open an Item 4K 13221 15 366 6 12855 9
HTML
Open an Item 4K 5041 7 186 3 4855 4
Plain Text
Open an Item with 18957 21 486 8 18471 13
Attachment 10K
HTML
Open an Item with 16246 18 426 7 15820 11
Attachment 10K
Plain Text
Open an Item with 79315 85 1996 33 77319 52
Attachment 50K
HTML
Open an Item with 77906 82 1866 31 76040 51
Attachment 50K
Plain Text
Open an Item with 153883 165 3906 65 149977 100
Attachment 100K
HTML
Open an Item with 154336 166 3966 66 150370 100
Attachment 100K
Plain Text

Open an Item with 748149 766 16986 283 731163 483
Attachment 500K
HTML
Open an Item with 764657 820 19626 327 745031 493
Attachment 500K
Plain Text
Open an Item with 1496056 1610 38572 642 1457484 968
Attachment 1000K
HTML
Open an Item with 1526516 1632 38886 648 1487630 984
Attachment 1000K
Plain Text
In general, less traffic is generated after the initial connection. During the initial
connection, the client requests and receives appropriate permissions, connects to
the NNTP server, and downloads the newsgroups list. The client keeps the list in a
local cache. In Outlook Express, the newsgroup list only shows the newsgroup
name and its description.
Graph 44 : Public Folder / Outlook Express / NNTP - Send Item without
16 000
14 000
12 000
10 000
Bytes
8 000
6 000
4 000
2 000
HTM L 5 490 8 383 14 332

Plain Text 2 831 3 906 6 169
HTML publications are still larger than Plain Text publications; however, sending
attachments is not format–sensitive.
108
Graph 45 : Public Folder / Outlook Express / NNTP - Send Item with attachment
1 600 000
1 400 000
1 200 000
1 000 000
Bytes
800 000
600 000
400 000
200 000
HTM L 20 002 80 363 154 802 750 406 1 494 197

Plain Text 17 297 79 028 155 249 764 598 1 527 879
The small gap is due to the 1KB text conversion. Sending attachments is not format
sensitive.
Graph 46 : Public Folder / Outlook Express / NNTP - Read Item without

14 000
12 000
10 000
8 000
Bytes
6 000
4 000
2 000
HTM L 4 415 7 423 13 221

Plain Text 1 801 2 822 5 041
Analysis on reading publications is similar to the sending publications analysis:

HTML format is bigger.
Graph 47 : Public Folder / Outlook Express / NNTP - Read Item with attachment
1 600 000
1 400 000
1 200 000
1 000 000
Bytes
800 000
600 000
400 000
200 000
Off ice Of fice Of fice Off ice Of fice
HTM L 18 957 79 315 153 883 748 149 1 496 056

Plain Text 16 246 77 906 154 336 764 657 1 526 516
Analysis on reading publications with attachment is similar to the sending

publications analysis: HTML format is bigger.
8.6. Netscape Messenger NNTP Results

PF1: Open “All 1307 17 640 10 667 7
Public Folders”
with 1 public
folder (1st time)
PF2: Open “All 392 5 198 3 194 2
Public Folders”
with 1 public
folder (2nd time)
PF5: Open “All 1013 11 380 6 633 5
Public Folders”
with 5 public
folder (1st time)
PF6: Open “All 349 3 120 2 229 1
Public Folders”
with 5 public
folder (2nd time)
110

PF10: Open “All 1159 11 380 6 779 5
Public Folders”
with 10 public
folder (1st time)
PF11: Open “All 495 3 120 2 375 1
Public Folders”
with 10 public
folder (2nd time)
Send an Item 1K 2633 10 2309 6 324 4
HTML
Send an Item 1K 2211 9 1887 5 324 4
Plain Text
Send an Item 2K 4509 13 4131 8 378 5
HTML
Send an Item 2K 3453 13 3021 7 432 6
Plain Text
Send an Item 4K 7744 19 7258 12 486 7
HTML
Send an Item 4K 5604 15 5172 9 432 6
Plain Text
Send an Item 17506 31 16804 20 702 11
with Attachment
10K HTML
Send an Item 25589 59 19192 34 6397 25
with Attachment
10K Plain Text
Send an Item 78809 114 76379 71 2430 43
with Attachment
50K HTML
Send an Item 78278 111 75902 69 2376 42
with Attachment
50K Plain Text
Send an Item 154104 209 149676 129 4428 80
with Attachment
100K HTML
Send an Item 153629 207 149309 129 4320 78
with Attachment
100K Plain Text
Send an Item 753963 923 735387 581 18576 342
with Attachment
500K HTML
Send an Item 753162 915 733938 561 19224 354
with Attachment
500K Plain Text

Send an Item 1503071 1815 1466567 1141 36504 674
with Attachment
1000K HTML
Send an Item 1496601 1702 1460475 1035 36126 667
with Attachment
1000K Plain
Text
Read an Item 1K 2105 4 126 2 1979 2
HTML
Read an Item 1K 1737 4 126 2 1611 2
Plain Text
Read an Item 2K 3933 6 186 3 3747 3
HTML
Read an Item 2K 2763 4 126 2 2637 2
Plain Text
Read an Item 4K 7006 9 246 4 6760 5
HTML
Read an Item 4K 4974 7 186 3 4788 4
Plain Text
Read an Item 16690 19 486 8 16204 11
with Attachment
10K HTML
Read an Item 16364 20 486 8 15878 12
with Attachment
10K Plain Text
Read an Item 76997 81 1866 31 75131 50
with Attachment
50K HTML
Read an Item 76629 81 1866 31 74763 50
with Attachment
50K Plain Text
Read an Item 151435 157 3546 59 147889 98
with Attachment
100K HTML
Read an Item 151067 157 3546 59 147521 98
with Attachment
100K Plain Text
Read an Item 747077 767 17046 284 729981 483
with Attachment
500K HTML
Read an Item 746545 765 16986 283 729559 482
with Attachment
500K Plain Text
112

Read an Item 1490688 1527 33846 564 1456842 963
with Attachment
1000K HTML
Read an Item 1490206 1525 33786 563 1456420 962
with Attachment
1000K Plain
Text
Usually, less traffic is generated after the initial connection. During the initial
connection, the client requests and receives appropriate permissions, connects to
the NNTP server, and downloads the newsgroups list. The client keeps the list in a
local cache. With Netscape Messenger, the list shows the newsgroup name, its
description, and the number of unread posts. This explains the difference in
network traffic consumed by Outlook Express and Netscape.
Graph 48 : Public Folder / Netscape Messenger / NNTP - Send item without
8 000
7 000
6 000
5 000
Bytes
4 000
3 000
2 000
1 000
HTM L 2 633 4 509 7 744

Plain Text 2 211 3 453 5 604
The gap is still there and lesser than the observed gap with Outlook Express.
Graph 49 : Public Folder / Netscape Messenger / NNTP - Send item with attachment
1 600 000
1 400 000
1 200 000
1 000 000
Bytes
800 000
600 000
400 000
200 000
0 1K t ext w/ 10K 1K t ext w/ 50K 1K t ext w/ 100K 1K text w/ 500K 1K t ext w/
Off ice Off ice Of fice Off ice 1000K Of fice
HTM L 17 506 78 809 154 104 753 963 1 503 071

Plain Text 25 589 78 278 153 629 753 162 1 496 601
The small gap is due to the 1KB text conversion. Sending attachments is not format
sensitive.
Graph 50 : Public Folder / Netscape Messenger / NNTP - Read item without

8 000
7 000
6 000
5 000
Bytes
4 000
3 000
2 000
1 000
HTM L 2 105 3 933 7 006

Plain Text 1 737 2 763 4 974
The gap is still there and lesser than the observed gap with Outlook Express.
114
Graph 51 : Public Folder / Netscape Messenger / NNTP - Send Item with attachment
1 600 000
1 400 000
1 200 000
1 000 000
Bytes
800 000
600 000
400 000
200 000
0 1K text w/ 10K 1K text w/ 50K 1K text w/ 100K 1K text w/ 500K 1K text w/ 1000K
Office Office Office Office Office
HTM L 16 690 76 997 151 435 747 027 1 490 688

Plain Text 16 364 76 629 151 067 746 545 1 490 206
Surprisingly, there almost no gap between formats.
8.7. OWA results
Public folders are available only via the Folders tab in the Outlook bar. The
concept of Favorites does not exist so there is no All Public Folders root folder ;
you have directly Public folders which contains all the public folders hierarchy.

Click on Folders in 14860 24 7608 13 7252 11
Outlook bar
Open a empty public 25217 42 6895 20 18322 22
folder
Click on Folders Bar in 3807 8 2556 5 1251 3
Outlook Bar with one
public folder
Outlook Bar with five
public folders

Outlook Bar with ten
public folders
New Post (1st time) 25143 30 3163 11 21980 19
Post a 1KB item 5557 13 4200 8 1357 5
Open a 1KB item 9778 11 513 4 9265 7
The whole hierarchy (first level of Public Folders) is downloaded when the user
clicks the Folders tab in the Outlook bar. Thus, no further traffic is generated when
the user opens the public folders tree or show the Folders tab (you need to refresh
the URL via Internet Explorer to refresh the hierarchy).
Deleting an public folder item is similar to delete message item.
8.8. Global results
Once connection and hierarchy listings have been established, less traffic occurs
during similar queries in the same session.
Graph 52 : Public Folder- Send item in HTML format without attachment
(comparative : Outlook Express, Netscape)
16 000
14 000
12 000
10 000
Bytes
8 000
6 000
4 000
2 000
Outlook Express 5 490 8 383 14 332

Netscape 2 633 4 509 7 744
116
Outlook Express generates more traffic than Netscape Messenger in posting HTML
publications, regardless of the size of the messages being handled. However, there
is a small difference in network traffic generated by Outlook Express and Netscape
for posting HTML publications with attachments. Outlook Express needs, on
average, 2 KB more bandwidth than Netscape to get a publication with an
attachment.
Graph 53 : Public Folder- Send item in HTML format with attachment (comparative :
Outlook Express, Netscape)
1 600 000
1 400 000
1 200 000
1 000 000
Bytes
800 000
600 000
400 000
200 000
0 1K t ext w/ 10K 1K text w/ 50K 1K text w/ 100K 1K text w/ 500K 1K t ext w/ 1000K
Off ice Off ice Office Office Office
Outlook Express 20 002 80 363 154 802 750 406 1 494 197
Netscape 17 506 78 809 154 104 753 963 1 503 071
Very small gap between Outlook Express and Netscape in posting HTML
publications with attachment.
Graph 54 : Public Folder- Send item in Plain Text format without attachment
(comparative : Outlook Express, Netscape)
7 000
6 000
5 000
4 000
Bytes
3 000
2 000
1 000
Out look Express 2 831 3 906 6 169

Net scape 2 211 3 453 5 604
Smaller gap between Outlook Express and Netscape in posting Plain Text
publications with no attachment.
Graph 55 : Public Folder- Send item in Plain Text format with attachment
(comparative: Outlook Express, Netscape)
1 600 000
1 400 000
1 200 000
Bytes 1 000 000
800 000
600 000
400 000
200 000
0 1K text w/ 10K 1K text w/ 50K 1K t ext w/ 100K 1K text w/ 500K 1K text w/ 1000K
Office Off ice Of fice Office Off ice
Outlook Express 17 297 80 363 155 249 764 598 1 527 879
Netscape 25 589 78 278 153 629 753 162 1 496 601
Very small gap between Outlook Express and Netscape in posting Plain Text
publications with attachment.
Graph 56 : Public Folder- Read item in HTML format without attachment

14 000
12 000
10 000
8 000
Bytes
6 000
4 000
2 000

Netscape 2 105 3 933 7 006
Whatever the size, Outlook Express generates more traffic than Netscape
Messenger in reading HTML publications.
118
Graph 57 : Public Folder- Read item in HTML format with attachment (comparative:
Outlook Express, Netscape)
1 600 000
1 400 000
1 200 000
1 000 000
Bytes
800 000
600 000
400 000
200 000
Outlook Express 18 957 79 315 153 883 748 149 1 496 056
Netscape 16 690 76 997 151 435 747 027 1 490 688
Public Folder- Read item in HTML format with attachment (comparative: Outlook
Express, Netscape)
Graph 58 : Public Folder- Read item in Plain Text format without attachment
6 000
5 000
4 000
Bytes
3 000
2 000
1 000

Netscape 1 737 2 763 4 974
No gap between Outlook Express and Netscape.

Graph 59 : Public Folder- Read item in Plain Text format with attachment
1 600 000
1 400 000
1 200 000
1 000 000
Bytes
800 000
600 000
400 000
200 000
Off ice Of fice Of fice Off ice Off ice
Out look Express 16 246 77 906 154 336 764 657 1 526 516
Net scape 16 364 76 629 151 067 746 545 1 490 206
The difference in network traffic is proportional to attachment size. With a 50–KB

attachment, there is difference of 1 KB. While with a 1–MB attachment, there is
gap of 36 KB. Outlook Express generates more traffic than Netscape when
downloading plain text publications with attachments.
120
9. Outlook 2000 with Terminal

Services
9.1. Introduction
Terminal Services, running on a computer running Windows 2000, enables all

client application execution, data processing, and storage to occur on the server.
Applications appear on the user’s device by means of terminal emulator software,
which can run on a variety of client hardware devices: a personal computer, a
Windows-based terminal, or even a Windows CE-based handheld device. The term
Windows-based terminal broadly refers to a class of thin client terminal devices,
such as Terminal Services, that can gain access to servers running a multi-user
Windows operating system.
Users can gain access to Terminal Services over any TCP/IP connection, including
Remote Access Ethernet, the Internet, wireless, WAN, or virtual private network
(VPN). The user experience is only limited by the characteristics of the weakest
link in the connection (for example, hardware capabilities), and the security of the
link is governed by the TCP/IP deployment in the data center.
With Terminal Services, the terminal emulation software sends keystrokes and
mouse movements to the server. Terminal Services manipulates data locally and
then passes back the display. This approach allows remote control of servers and
centralized application management, minimizing network bandwidth requirements
between the server and client.
Some companies may choose to provide access to Outlook 2000 through terminal
emulation software. This section evaluates the impact on network traffic, when
using Outlook 2000 through the 32-bit Terminal Services client, on a computer
running Windows 2000 Server.
9.2. Test Details
The Terminal Services client sends keystrokes and mouse movements to the server.
For that reason, the tests provide differing results because it is impossible to ensure
exact mouse movements. These tests were performed by moving the mouse very
carefully, generating as little additional traffic as possible. The traffic captured in
these tests is only that between the client and the Terminal server. During a
Terminal Services session, no traffic occurs between the Terminal client computer
and the Exchange server.
Connect
Captures the traffic generated by initial connection to the server. The capture was
stopped when the logon dialog box appeared.
CONNECT: open a Terminal Services session
1. Start Terminal Services client
2. Click Connect.
TS Logon
Captures the traffic generated by opening a session on the server. The test stopped
when the desktop appeared.
TSLOGON: open a session on the Terminal server
1. Enter user name and password.
2. Press ENTER.
TS Logoff
Captures the traffic resulting from closing a session on the server.
TSLOGOFF: close the Terminal Services session
1. Quit all running applications.
2. Click Start, click Shutdown.
3. In the Shut down Windows box, select Log off current_user.
Outlook Logon
Captures the traffic generated by opening an Outlook session on the Terminal
server. The Outlook preview pane is open.
LOGON1: open Outlook containing no messages
1. Create a profile in an empty mailbox.
2. Double-click Outlook icon on the desktop.
122
LOGON2: open Outlook containing five messages

1. Put five unread messages in your Inbox.
2. Quit Outlook.
LOGON3: open Outlook containing ten messages

1. Put ten unread messages in your Inbox.
2. Quit Outlook.
Outlook Logoff
Captures the traffic resulting from closing the Outlook session.
LOGOFF: logoff Outlook
1. Close all Outlook windows, except the main window.
2. Press ALT+F4.
Messages actions
Captures the traffic resulting from the last step in each of the following tests.
MSG1: Open Inbox folder containing no messages
1. Open Outlook.
2. Open Deleted Items folder.
3. Click Inbox on the Outlook bar.
MSG2: Display the new message form.

1. Click File, point to New, and then click Mail Message.
MSG3: New message form with CTRL+N

1. From the Inbox view, press CTRL+N.
MSG4: New message form with CTRL+N after x attempts

1. From the Inbox view, press CTRL+N.
MSG5: Send a message

1. Click File, point to New, and then click Mail Message.
2. Fill the To and Subject boxes.
3. Click Send.
MSG6: Open Outlook Today containing one appointment and one task
1. Open Outlook.
2. Click Outlook Today on the Outlook bar.
Calendar actions
CAL1: Open Calendar folder with no entry
1. Open Outlook.
2. Click Calendar on the Outlook bar.
CAL2: New appointment form

1. From the Calendar view, click File, point to New, and then click
Appointment.
CAL3: New appointment form with CTRL+N

1. From the Calendar view, press CTRL+N.
CAL4: New appointment form with CTRL+N after x attempts

1. From the Calendar view, press CTRL+N.
CAL5: Save and close an appointment

1. From the Calendar view, click File, point to New, and then click
Appointment.
2. Fill the necessary boxes.
Contacts actions
CTC1: Open Contacts folder containing no entries
1. Open Outlook.
2. Click Contacts on the Outlook bar.
CTC2: Display the new contact form

1. From the Contacts view, click File, point to New, and then click Contact.
CTC3: New Contact form with CTRL+N

1. From the Contacts view, press CTRL+N.
CTC4: New Contact form with CTRL+N after x attempts

1. From the Contacts view, press CTRL+N.
124
CTC5: Save and close a contact

1. From the Contacts view, click File, point to New, and then click Contact.
Tasks actions
Captures the traffic resulting from the last step in each following test.
TSK1: Open Tasks folder containing no entries
1. Open Outlook.
2. Click Tasks on the Outlook bar.
TSK2: Display the new task form

1. From the Tasks view, click File, point to New, and then click Task.
TSK3: Display the new Task form with CTRL+N

1. From the Tasks view, press CTRL+N.
TSK4: New Task form with CTRL+N after x attempts

1. From the Tasks view, press CTRL+N.
TSK5: Save and close a task

1. From the Tasks view, click File, point to New, and then click Task.
Notes actions
Captures the traffic resulting from the last step in each following test.
NOT1: Open Notes folder containing no entries
1. Open Outlook.
2. Click Notes on the Outlook bar.
NOT2: Display the new Note form

1. From the Notes view, click File, point to New, and then click Note.
NOT3: New Note form with CTRL+N

1. From the Notes view, press CTRL+N.
NOT4: New Note form with CTRL+N after x attempts

1. From the Notes view, press CTRL+N.
NOT3: Save and close a Note

1. From the Notes view, click File, point to New, and then click Note.
3. In the upper-right part of the screen, click Close.
Address Book actions

AR1: Check one name
2. In the To box, enter an alias that will resolve to a single name.
3. Press CTRL+K (Check Names).
AR2: Full check one name

2. In the To box, enter an alias that will resolve to a single name.
3. Press the button Check Names.
AL1: Display the address book with the To button

2. Click To.
AB1: Add one address in To

2. Click To.
3. Select one address, click To, and then click OK.
AD1: Show Details in Address Book

2. Click To.
3. Select one address.
9.3. Terminal Services Client Results
Client to
Received by Client to TS TS Server to Exchange 2000
Total Sent by Client Exchange 2000
Client server Client Server to Client
Server
Bytes Frames Bytes Frames Bytes Frames Bytes Frames Bytes Frames Bytes Frames Bytes Frames
Connect 18 024 69 4 795 33 13 229 36 4 795 33 0 0 13 229 36 0 0
Open TS session 13 812 46 1 321 18 12 491 28 1 321 18 0 0 12 491 28 0 0
Close TS session 13 910 79 3 569 38 10 341 41 2 200 29 1 369 9 9 074 34 1 307 7
Sent by Client Received by Client

Total
(to TS server) (from TS server)

LOGON1: with welcome
message 6728 82 2952 39 3776 43
LOGON2: no message 3170 42 1521 21 1649 21
LOGON3: 5 messages 4465 47 1733 23 2732 24
LOGON4: 10 messages 3907 53 1881 25 2026 28
Logoff 2501 27 1036 13 1465 14
MSG1: Open Inbox folder 2135 19 612 9 1523 10
MSG2: New Mail Message
(mouse) 7615 39 1542 19 6073 20
with CTRL+N 7507 20 564 9 6943 11
with CTRL+N after x
attempts in the session 1266 15 444 7 822 8
MSG5: Send Mail message 4432 35 1387 19 3045 16
MSG6: Open Outlook
Today 6040 39 1475 19 4565 20
CAL1: Open Calendar folder 6480 31 1011 14 5469 17
CAL2: New Appointment
(mouse) 10026 33 1308 16 8718 17
with CTRL+N 8860 20 504 8 8356 12
with CTRL+N after x
attempts in the session 1266 15 444 7 822 8
CAL5: Save & Close
appointment 3600 29 1265 15 2335 14
CTC1: Open Contacts folder 3314 21 930 11 2384 10
CTC2: New Contact 10669 39 1470 19 9199 20
CTC3: New Contact with
CTRL+N 9194 19 504 8 8690 11
CTC4: New Contact with
CTRL+N after x attempts in
the session 1595 15 444 7 1151 8
CTC5: Save & Close
Contact 4141 32 1301 17 2840 15
TSK1: Open Tasks folder 6521 28 1174 14 5347 14
TSK2: New Tasks 8504 31 1152 14 7352 17
TSK3: New Task with
CTRL+N 7032 17 444 7 6588 10
TSK4: New Task with
the session 1659 17 444 7 1215 10
TSK5: Save & Close Task 3083 27 1114 13 1969 14
NOT1: Open Notes folder 4510 37 1460 19 3050 18
NOT2: New Note 3170 30 1238 16 1932 14
NOT3: New Note with
CTRL+N 1573 12 384 6 1189 6
128
Sent by Client Received by Client

Total
(to TS server) (from TS server)

NOT4: New Note with
the session 945 12 384 6 561 6
NOT5: Save & Close Note 3461 33 2006 18 1455 15
AR1: Check names 1244 15 444 7 800 8
AR2: Full check one name 4290 50 1620 25 2670 25
AL1: Show AB with To
button 2271 23 877 11 1394 12
AB1: Add on add to To field
from AB 4992 52 2282 27 2710 25
AD1: Details from AB 3299 32 1574 18 1725 14
9.4. Terminal Services Measurement Analysis
These tests were done with an 800 x 600 pixel resolution for the Terminal Services
session. Other captures were done on the tests “Connect” and “Open Terminal
Services session” with a lower resolution (640 x 480) and higher resolution (1024 x
768). These new captures had the same traffic volume (bytes and frames).
To reduce traffic during Terminal Services sessions, use keyboards shortcuts as
much as possible. Mouse events are sent based on a regular frequency and
increased network traffic. In addition, when configuring a Terminal Services
connection, select the option to enable data compression; this will also decrease
traffic.
10. Web Storage System
10.1. Introduction
The Microsoft Web Storage System is a database for messaging, collaboration, rich
document storage, and Web-enabled applications. The Web Storage System can be
accessed by a wide range of client software, including:
• Microsoft Outlook 97, Outlook 98, and
Outlook 2000 messaging and
collaboration clients
• Outlook Express and any e-mail or
newsgroup client that supports
SMTP/POP3, IMAP4, or NNTP
• Microsoft FrontPage
• Microsoft Office 2000
• Windows Explorer
• Web Folders (included with Internet
Explorer 5, Office 2000, and
Windows 2000)
• Any Web browser
• The MS-DOS prompt
• Any 32-bit application for Windows
Upon installation, the Web Storage System is mapped to the M drive on the
Exchange 2000 server, and accessed in the same way as the existing Windows file
system. The administrator can share this virtual drive to give access to users who
can then access their mailboxes and public folders as with any file server. You can
access through Uniform Naming Convention (UNC) or a mapped drive letter.
130
Seven test categories evaluated Web Storage System traffic:

• Tools/map network drive Connect to the network share and map a network
drive with Windows 2000.
• Net use Map a network drive with the command line, NET USE.
• Logoff Disconnect from the network share.
• Create folder Create a new folder within your mailbox.
• Copy an item: x–KB (10–KB, 50–KB, 100–KB, 500–KB, and 1,000–KB) Copy
a file into one of the mailbox folders.
• Open an item: x–KB (10–KB, 50–KB, 100–KB, 500–KB, and 1,000–
KB) Open a file in one of the mailbox folders.
• Delete an item: x–KB (10–KB, 50–KB, 100–KB, 500–KB, and 1,000–
KB) Delete a file in one of the mailbox folders.
10.3. Log On and Log Off and User Traffic Results
Received by
Client
Fram Frame Fram
Bytes Bytes Bytes
es s es
With UI via Tools/Map 7112 44 4502 24 2610 20
network drive
With command line NET USE 3768 18 2588 10 1180 8
Logoff 0 0 0 0 0 0
10.4. User traffic

Create Folder 18391 94 8465 48 9926 46
Copy a 10k 65366 249 29520 127 35846 122
document
Copy a 50k 96456 283 69889 150 26567 133
document
Copy a 100k 175090 343 122554 185 52536 158
document
Copy a 500k 598897 817 551546 493 47351 324
document
Copy a 1000k 1225183 1314 1074416 801 150767 513
document
Read a 10k 17679 43 2527 19 15152 24
document
Read a 50k 58702 77 3001 25 55701 52
document
Read a 100k 112399 120 3544 33 108855 87
document
Read a 500k 541352 460 6993 85 534359 375
document
Read a 1000k 1077048 883 11415 151 1065633 732
document
Delete a 10k 15397 84 7347 43 8050 41
document
Delete a 50k 15397 84 7347 43 8050 41
document
Delete a 100k 15409 84 7359 43 8050 41
document
Delete a 500k 15409 84 7359 43 8050 41
document
Delete a 1000k 15437 84 7371 43 8066 41
document
132
10.5. Web Storage System Measurement Analysis
Theuser traffic results are similar to the traffic generated when you work with a
standard shared directory on a file server. Exchange 2000 Server does not generate
network traffic; all processing is performed on the server.
If you compare the user traffic results with sending or opening such attachments
with a MAPI client, you will not see big difference. The Web Storage System is not
a particularly more or less efficient way to work with your mailbox; it is just a
convenient method for some users. It makes mailbox access easier because
Collaboration Data Objects (CDO) is not required. You can now run batch files on
your mailbox.
11. Instant Messaging
11.1. Introduction
Instant Messaging consists of immediate, text-based messages that you can send to
other users on a computer network. Unlike e-mail messages, these messages post to
the other user’s screen, providing the basis for new forms of collaboration. Instant
Messaging has become a wide-scale communication phenomenon for Internet users
and is poised to play a significant role as a business tool for organizations of all
sizes. Exchange 2000 includes an Instant Messaging service built on a secure,
standards-oriented architecture ideally suited for both enterprise deployment and
deployments across the Internet for business-to-business communication. The
client software for instant messaging in Exchange 2000 is the MSN Messenger
client.
Presence information, closely related to Instant Messaging, enables one computer
user to see whether another user is logged onto a network, corporate LAN, or the
Internet. Exchange 2000 provides complete support for presence information.

• Logon with 0, 1 and 10 contacts Open an Instant Messaging session with no
contacts, then 1 contact, then 10 configured contacts.
• Add a contact Add a contact in the contacts list.
• Delete a contact Delete a contact from the contacts list.
• Change status Statuses include:
• Online
• Busy
• Be right back
• Away
• On the phone
• Out to lunch
• Appear offline
134
The traffic was tracked on the contact side, and on the user side (where the contact
is added).
• Send and receive messages This function is divided into three parts:
• Entering characters with keyboard
• Sending the message
• Receiving the message
11.3. Log On and Log Off

Logon without contact 31844 116 8858 61 22986 55
Logon with 1 contact 33439 125 9480 66 23959 59
Logon with 10 contacts 47682 206 15055 111 32627 95
Logoff without contact 14910 52 4055 27 10855 25
Logoff with 1 contact 16050 61 4573 32 11477 29
Logoff with 10 contacts 26237 142 9185 77 17052 65
11.4. User Traffic Results

Add one Contact 1594 9 622 5 972 4
Delete one Contact 1130 9 515 5 615 4
Change status to ONLINE 8519 29 2797 15 5722 14
Change status to BUSY 8511 29 2791 15 5720 14
Change status to BE 8531 29 2806 15 5725 14
RIGHT BACK
Change status to AWAY 8511 29 2791 15 5720 14

Change status to ON THE 8527 29 2803 15 5724 14
PHONE
Change status to OUT TO 8527 29 2803 15 5724 14
LUNCH
Change status to APPEAR 8016 29 2422 15 5594 14
OFFLINE
Contact change status to 1398 7 424 3 974 4
ONLINE
BUSY
BE RIGHT BACK
AWAY
ON THE PHONE
OUT TO LUNCH
APPEAR OFFLINE
Enter first, one character 2102 12 1366 6 736 6
(CASE1)
Enter first, at least 2 4204 24 2732 12 1472 12
characters (CASE2)
Send x characters 2102 12 1366 6 736 6
Receive first character 1622 7 424 3 1198 4
(CASE1)
Receive first, at least two 3244 14 848 6 2396 8
characters (CASE2)
Receive x characters 1622 7 424 3 1198 4
Press « Send » button 2131 12 1395 6 736 6
Receive the end of the 1651 8 424 4 1227 4
message
136
11.5. Global results
The Instant Messaging client generates network traffic when the status changes or
when messages are sent.
11.5.1. Status Change

When users change their status, information goes to the Exchange 2000 server,
which manages the Instant Messaging service. The server sends the status change
to all users who have that user in their contacts list. The new status is then reflected
in the contacts list.
All status changes do not generate the same network traffic volume. The status
change to “Away” can happen automatically when no activity is detected. This
feature is configured in the option Show me as Away when I’m inactive for x
minutes, available in the MSN Messenger client. In that case, when no activity is
detected for x minutes, the client switches to “Away” status, and notifies the server
of the change. When an activity is detected again, the client switches back to online
status.
11.5.2. Sending Messages

The amount of network traffic required for messages depends on touch speed. As a
message is created, if the first pause occurs after the first character is typed, the
sender exchanges 12 frames with the server, and the server exchanges seven frames
with the recipient.
If the first pause occurs after least two characters are typed, 24 frames exchange
between the server and the sender and 14 are exchanged between the server and the
recipient.
Afterwards, multiples of 12 frames are exchanged between sender and server, and
multiples of seven frames are exchanged between server and recipient. This traffic
occurs until the user clicks Send. In actuality, the recipient receives the message
content before sending occurs. This accounts for the rapidity of Instant Messaging.
Clicking Send creates 16 frames between the sender and the server, and 8 frames
between the server and the recipient.
A copy or paste into the message body does not generate any traffic.
12. Client Traffic Measurement

Conclusions
There are many things to learn from these tests.

• Outlook 2000 and Outlook 97 require approximately the same bandwidth.
• Logging on represents a large part of the traffic with MAPI clients.
• Outlook 2000 is more integrated with Windows 2000 because of the Outlook
capacity to query Active Directory.
• Of the different formats tested, HTML messages generate the most traffic, and
RTF messages generate the least.
• Attachments only download when users want to see them.
• Except for RTF messages, no compression occurs between clients and servers. The
full name is RTF-compressed, so messages are compressed between MAPI clients
and the server. Therefore, HTML (no compression) generates more network traffic
when messages are downloaded. Many third-party tools fix this critical issue for
remote users with low bandwidth connections.
• Among Internet protocols, POP3 generates the least network traffic. Netscape
Messenger seems to send HTML messages more efficiently than Outlook Express.
• The use of XML in Outlook Web Access significantly reduces traffic. Forms
design is cached. In some situations, Outlook Web Access should be considered
over Outlook 2000 with Terminal Services.
• Mouse movements generate excess network traffic when using Outlook 2000 with
Terminal Services. Using keyboard reduces dramatically the network traffic.
• Web folder offers a new and easy way to access to mailboxes or publics folders.
Because Web folder access generates the same traffic as common file share access,
it does not increase network traffic.
138
13. How client use Front-End /

Back-End topology ?
This section provides the basis of interactions between Exchange 2000 servers
within an organization. It covers these subjects:
• Network traffic between one front-end server and back-end servers
• Network traffic between two routing groups
• Network traffic between one Exchange 2000, Domain Controller and
Global Catalog.
Pour theses tests, a new had been prepared.
13.1. Test Lab Configuration
13.1.1. Lab design

The test lab is built with 7 Windows 2000 servers. These servers host Exchange
2000 Server and other Windows 2000 services like Domain controller, Global
Catalog and DNS.
Figure 7 : Network Topology
WINDOWS 2000 Server

Domain Controller
Global Catalog Server
BIRD01
WINDOWS 2000 Server

Member Server WINDOWS 2000 Professional
DNS Service Outlook 2000
BIRD02 Outlook Express 5.0
BIRD08 Internet Explorer 5.0
WINDOWS 2000 Server

Domain Controller
BIRD03

WINDOWS 2000 Server Outlook Express 5.0
Member Server Internet Explorer 5.0
BIRD09
BIRD04
WINDOWS 2000 Server

Member Server
BIRD05

WINDOWS 2000 Server
Network Monitor
Member Server
Exchange 2000 Server BIRD10
BIRD06
WINDOWS 2000 Server

Member Server
BIRD07
13.1.2. LAN design

The DNS root domain is called Europe.com. The DNS primary zone is managed by
the server BIRD02. This zone allows « Dynamic Updates » (Yes is set in the
General tab of this zone properties). Every server has a static IP address and refers
to the DNS to resolve names.
In order to not disturb captures and to run into caching phenomenon, the MAC and
IP addresses are preloaded on the client stations. It prevents to find sometimes
ARP or DNS queries in the captures.
The network is a Ethernet à 100 Mbs dedicated network.
140
Table 4 : Server configurations
Server IP address IP Subnet mask Role

BIRD01 10.0.0.1 255.0.0.0 Domain Controller
Global Catalog
BIRD02 10.0.0.2 255.0.0.0 Member Server
DNS
BIRD03 10.0.0.3 255.0.0.0 Domain Controller
(Back-End)
(Back-End)
(Front-End)
(Back-End)
BIRD08 10.0.0.8 255.0.0.0 Member workstation
13.1.3. Servers characteristics

each server.
Table 5 : Server characteristics

Software
BIRD01 Intel Pentium 128 Mb Windows 2000
III 500 Advanced Server
III 500 Advanced Server
BIRD03 Intel Pentium II 128 Mb Windows 2000
450 Advanced Server
BIRD04 4 x Intel 2 Gb Windows 2000 Exchange
Pentium III 550 Advanced Server 2000

Software
BIRD07 Intel Pentium II 128 Mb Windows 2000 Exchange
333 Advanced Server 2000
13.1.4. Clients characteristics

Table 6 : Client characteristics
Client CPU Memory OS Other
Software
BIRD08 Intel Pentium 128 Mb Windows 2000 Outlook
III 500 Professional 2000
III 500 Professional
BIRD10 Intel Pentium 128 Mb Windows 2000 Network
III 500 Professional Monitor
The network traffic captures are done with Microsoft Network Monitor 2.0. It is
the version 5.00.943 available in the Microsoft System Management Server 2.0,
Service Pack 2. This version works fine under Windows 2000.
13.1.5. Exchange 2000 configuration
All Exchange 2000 servers are installed with default settings.

142
Figure 8 : Exchange 2000 installed components
In order to study exchanges between servers, we create 2 routing groups :

• Paris, which includes the servers BIRD04, BIRD05 and BIRD06.
• Nantes, which includes the server BIRD07.
These 2 routing groups are linked with different connectors according the tests.
Figure 9 : Routing Groups Configuration

The BIRD06 server is set as a Front-End server. It will be used as mailbox server
for internet clients, even if the mailboxes are on one of the three other servers
(BIRD04, BIRD05, BIRD07): this is the front-end concept. You change a standard
server (back-end) to a front-end by checking the option « This is a front-end
server » option as shown below. After this change, you need to restart the services
W3SRV, POP and IMAP to apply changes.
Figure 10 : Front-End Switch
To not disturb captures, we have made adjustments on Exchange 2000 settings and
registry keys.
The Exchange 2000 server query Active Directory on a Global Catalog to get
information.
Configuration information are get by requesting the AD of a domain controller
(DC). We can force the queried DC by setting the following registry entry :
HKEY_LOCAL_MACHINE \ System \ CurrentControlSet \ Services \

MSExchangeDSAccess \ Instance0
Value name: ConfigDCHostName

Value type: REG_SZ
Value name: ConfigDCPortNumber

Value type: REG_DWORD
Value data: 0x389
On lab servers, we set the ConfigDCHostName registry entry to

BIRD03.EUROPE.COM.
You can also force the server used to get Address Book information. In that case, a GC
is used. You can set the GC name in the following registry entry :
144

MSExchangeDSAccess \ Profiles \ Default
Value name: UserGC1

Value type: REG_SZ
Value name: PortNumber

Value data: 0x3268 (the standard port number for a GC)
Value name: IsGC

Value data: 0x1 (Always set to 1 if the server specified
is a GC)
On lab servers, we set the UserGC1 registry entry to BIRD01.EUROPE.COM.
By default, Directory Access has a cache to keep last resolved information. This cache
allows to not stress the GC by always query it for the same information (sender,
recipients). The default lifetime is 10 minutes. The 2 following registry entries allow to
enable or disable the cache and to set the lifetime.

MSExchangeDSAccess
Value name: CachingEnabled

Value data: 0x1 (Enabled – Default) 0x2 (Disabled)

MSExchangeDSAccess \ Instance0
Value name: CacheTTL

Value data: 0x600 (seconds)
On lab servers, the cache is disabled, so we set the CachingEnabled entry to 2.
Last but not least, many process are running inside a Exchange 2000 to insure that
information is up to date. The « Recipient Update Service » is one of them. Its function
is to keep Address Lists up to date. By default, its schedule to run is set to « Always ».
Because, we do not create/remove users, we set the frequency to “Never” in the lab. We
need to set this value for both the Enterprise Configuration and for the Organization
level (Global Address List).
Figure 11 : Recipient Update Services View

146
14. Front-End / Back-End

Communications
Setting a Front-End / Back-End configuration with Exchange 2000 servers allows

Internet clients to know only the front-end. Then, the front-end routes the queries to the
right back-end server. There is no direct link between Internet clients and a back-end
server. The front-end is only design to managed all traffic from Internet clients : it
queries the directory (GC) to get information on sender (among them its home server)
and recipients (among them their home server) to route afterwards the query which can
be a SMTP message or any POP3/IMAP4 requests.
14.1. Message Sending
To understand the message sending process in a Front-End / Back-End configuration,

we sent a short message from TEST1@EUROPE.COM (his mailbox home server is
BIRD04) to the recipient TEST2@EUROPE.COM (his mailbox home server is
BIRD04). We use Outlook Express with a IMAP account on from the client station
BIRD08. The account is set to not copy sent items in « Sent items » folder.
Here are the frames which are generated by sending this message :
The client initializes the TCP connection with the front-end server on port 25
(SMTP)
1 11.576646 BIRD08 Clt BIRD06 FE X2000 TCP ....S., len: 0, seq:2361275243-2361275243, ack BIRD08 Clt
BIRD06 FE IP
2 11.576646 BIRD06 FE X2000 BIRD08 Clt TCP .A..S., len: 0, seq:1769021499-1769021499, ack BIRD06
FE BIRD08 Clt IP
3 11.576646 BIRD08 Clt BIRD06 FE X2000 TCP .A...., len: 0, seq:2361275244-2361275244, ack BIRD08 Clt
BIRD06 FE IP
The SMTP service on the front-end server responds. The SMTP dialog can
begin
4 11.576646 BIRD06 FE X2000 BIRD08 Clt SMTP Rsp: Service ready, 116 bytes BIRD06 FE
BIRD08 Clt IP
The client presents itself to the front-end server

5 11.576646 BIRD08 Clt BIRD06 FE X2000 SMTP Cmd: Hello, host identifier, 13 bytes BIRD08 Clt
BIRD06 FE IP
6 11.576646 BIRD06 FE X2000 BIRD08 Clt SMTP Rsp: Requested mail action okay, completed, 40 by
BIRD06 FE BIRD08 Clt IP
The sender name is sent to the front-end server

7 11.576646 BIRD08 Clt BIRD06 FE X2000 SMTP Cmd: Mail from <test1@europe.com>, 31 bytes
BIRD08 Clt BIRD06 FE IP
The recipient name is sent to the front-end server

9 11.576646 BIRD08 Clt BIRD06 FE X2000 SMTP Cmd: Recipient <test2@europe.com>, 29 bytes
The message body is sent to the front-end server

11 11.576646 BIRD08 Clt BIRD06 FE X2000 SMTP Cmd: Data: Mail data to follow, 6 bytes BIRD08
Clt BIRD06 FE IP
12 11.576646 BIRD06 FE X2000 BIRD08 Clt SMTP Rsp: Enter mail ..., 46 bytes BIRD06 FE
BIRD08 Clt IP
13 11.576646 BIRD08 Clt BIRD06 FE X2000 SMTP Data - continued from frame 11, 1202 bytes
14 11.696819 BIRD06 FE X2000 BIRD08 Clt TCP .A...., len: 0, seq:1769021772-1769021772, ack
15 11.696819 BIRD08 Clt BIRD06 FE X2000 SMTP Data - continued from frame 13, 6 bytes BIRD08
Clt BIRD06 FE IP
The client quits the connection to the front-end server

17 11.716848 BIRD08 Clt BIRD06 FE X2000 SMTP Cmd: Quit, losing connection, 6 bytes BIRD08
Clt BIRD06 FE IP
The front-end server acknowledges SMTP connection closing

18 11.716848 BIRD06 FE X2000 BIRD08 Clt SMTP Rsp: Service closing transmission channel, 66 byt
The TCP connection is closed between the client and the front-end server
19 11.716848 BIRD06 FE X2000 BIRD08 Clt TCP .A...F, len: 0, seq:1769021919-1769021919, ack
20 11.716848 BIRD08 Clt BIRD06 FE X2000 TCP .A...., len: 0, seq:2361276536-2361276536, ack
21 11.716848 BIRD08 Clt BIRD06 FE X2000 TCP .A...F, len: 0, seq:2361276536-2361276536, ack
148
The front-end server sends a LDAP query to BIRD01 (GC) on port 3268
(specific GC port)
23 11.716848 BIRD06 FE X2000 BIRD01 GC LDAP ProtocolOp: SearchRequest (3) BIRD06 FE
BIRD01 IP
+ Frame: Base frame properties
+ ETHERNET: ETYPE = 0x0800 : Protocol = IP: DOD Internet Protocol
+ IP: ID = 0x65AE; Proto = TCP; Len: 1326
+ TCP: .AP..., len: 1286, seq:1730121002-1730122288, ack: 33021289, win:16762, src: 1882
dst: 3268
+ LDAP: ProtocolOp: SearchRequest (3)
It searches sender (test1@europe.com) and recipient

(test2@europe.com)
LDAP: Filter Type = Or
LDAP: Filter Type = Equality Match
LDAP: Attribute Type = ProxyAddresses
LDAP: Attribute Value = smtp:test1@europe.com
The requested fields are:

DistinguishedName, forwardingAddress, objectClass, mail, member, ProxyAddresses,
textEncodedORAddress, legacyExchangeDN, objectGUID, altRecipient, DeliverAndRedirect,
HideDLMembership, delivContLength, SubmissionContLength, TargetAddress, HomeMDB,
msExchHomeServerName, AuthOrig, UnauthOrig, DLMemSubmitPerms, DLMemRejectPerms,
msExchRecipLimit, msExchMemberFilter, msExchMemberBaseDN, ReportToOriginator,
ReportToOwner, managedBy, cn, MapiRecipient, InternetEncoding, HomeMTA,
msExchExpansionServerName, msExchMailboxGuid, OOFReplyToOriginator, modifyTimestamp,
Heuristics, delivExtContTypes, DelivEITs, EncapsulationMethod, owner, SecurityProtocol,
objectCategory, deliveryMechanism, autoReply, MDBStorageQuota, MDBOverQuotaLimit,
MDBUseDefaults, AuthOrigBL, UnauthOrigBL, DLMemRejectPermsBL, DLMemSubmitPermsBL,
ProtocolSettings, deletedItemFlags, GarbageCollPeriod, mDBOverHardQuotaLimit,
showInAddressBook, ntSecurityDescriptor, displayName, objectSID, sn, givenName,
mailNickName, DisplayNamePrintable, adminDisplayName, msExchHomePublicMDB,
msExchMasterAccountSid, msExchMailboxSecurityDescriptor, userAccountControl, notes,
msExchQueryBaseDN
GC responds to front-end with many frames, then connection is closed

24 11.726862 BIRD01 GC BIRD06 FE X2000 LDAP ProtocolOp: SearchResponse (4) BIRD01
BIRD06 FE IP
25 11.726862 BIRD01 GC BIRD06 FE X2000 TCP .A...., len: 1460, seq: 33022749-33024209, ack:1
BIRD01 BIRD06 FE IP
BIRD01 BIRD06 FE IP
BIRD01 BIRD06 FE IP
BIRD01 BIRD06 FE IP
29 11.726862 BIRD06 FE X2000 BIRD01 GC TCP .A...., len: 0, seq:1730122288-1730122288, ack
BIRD06 FE BIRD01 IP
30 11.726862 BIRD01 GC BIRD06 FE X2000 TCP .AP..., len: 33, seq: 33028589-33028622, ack:1
BIRD01 BIRD06 FE IP
31 11.726862 BIRD06 FE X2000 BIRD01 GC TCP .A...., len: 0, seq:1730122288-1730122288, ack
BIRD06 FE BIRD01 IP
Front-end server opens a TCP connection with recipient back-end server on

port 25 (SMTP)
32 12.698259 BIRD06 FE X2000 BIRD04 BE X2000 TCP ....S., len: 0, seq:1769382124-1769382124, ack
BIRD06 FE BIRD04 BE X2000 IP
33 12.708273 BIRD04 BE X2000 BIRD06 FE X2000 TCP .A..S., len: 0, seq:1840745563-1840745563, ack
BIRD04 BE X2000 BIRD06 FE IP
34 12.708273 BIRD06 FE X2000 BIRD04 BE X2000 TCP .A...., len: 0, seq:1769382125-1769382125, ack
SMTP service on back-end server responds. The SMTP dialog can begin
35 12.708273 BIRD04 BE X2000 BIRD06 FE X2000 SMTP Rsp: Service ready, 116 bytes BIRD04 BE
X2000 BIRD06 FE IP
Back-end server authenticates front-end server

36 12.708273 BIRD06 FE X2000 BIRD04 BE X2000 SMTP Data, 24 bytes BIRD06 FE BIRD04 BE
X2000 IP
37 12.708273 BIRD04 BE X2000 BIRD06 FE X2000 SMTP Rsp: Requested mail action okay, completed,
231 b BIRD04 BE X2000 BIRD06 FE IP
38 12.708273 BIRD06 FE X2000 BIRD04 BE X2000 SMTP Data - continued from frame 36, 142 bytes
39 12.708273 BIRD04 BE X2000 BIRD06 FE X2000 SMTP Rsp: Unknown response type, 68 bytes
The sender name is sent to the back-end server

40 12.708273 BIRD06 FE X2000 BIRD04 BE X2000 SMTP Cmd: Mail from <test1@europe.com>, 40 bytes
41 12.708273 BIRD04 BE X2000 BIRD06 FE X2000 SMTP Rsp: Requested mail action okay, completed, 41
by BIRD04 BE X2000 BIRD06 FE IP
The recipient name is sent to the back-end server

42 12.708273 BIRD06 FE X2000 BIRD04 BE X2000 SMTP Cmd: Recipient <test2@europe.com>, 28 bytes
The message body is sent to the back-end server

45 12.718288 BIRD04 BE X2000 BIRD06 FE X2000 SMTP Rsp: Enter mail ..., 22 bytes BIRD04 BE
X2000 BIRD06 FE IP
49 12.898547 BIRD04 BE X2000 BIRD06 FE X2000 TCP .A...., len: 0, seq:1840746087-1840746087, ack
Front-end server quits the connection to the back-end server

52 12.918576 BIRD06 FE X2000 BIRD04 BE X2000 SMTP Cmd: Quit, losing connection, 6 bytes
Back-end server acknowledges SMTP connection closing

53 12.918576 BIRD04 BE X2000 BIRD06 FE X2000 SMTP Rsp: Service closing transmission channel, 66
byt BIRD04 BE X2000 BIRD06 FE IP
The TCP connection is closed between the front-end and the back-end server
54 12.918576 BIRD06 FE X2000 BIRD04 BE X2000 TCP .A...F, len: 0, seq:1769384832-1769384832, ack
55 12.918576 BIRD04 BE X2000 BIRD06 FE X2000 TCP .A...F, len: 0, seq:1840746234-1840746234, ack
150
The following drawing summarizes the chronology of exchanged frames between client,
front-end and back-end server.
Figure 12 : FE/BE Message Sending Traffic Frames Overview
Depends on message
body length
1 3 5 7 9 11 ... 16 17 20 21
BIRD08 BIRD06
Client 2 4 6 8 10 11 ... 16 18 19 22 Front-End
32 33
23 24
29 25 34 35
31 26
36 37
27
28 38 39
30
40 41
42 43
BIRD01
GC
44 45
...
45
Depends on message 51
body length ...
51 53
52 55
54 56
57
BIRD04
Back-End
14.1.1. Analysis
An Internet client only communicates with the front-end server. The front-end server
send LDAP queries to GC to know who are the sender and recipient, and especially to
know where is the home server of the recipient. Then, it re-sends the SMTP frames, it
received, to the recipient home server. It acts as a router.
If we enable “Save copy of sent messages in ‘Sent Items ‘folder”, the front-end server
send the message to the sender mailbox home server.
14.2. Message reading
To understand the message reading process in a Front-End / Back-End configuration,

we read a message from anyone in the mailbox of TEST1@EUROPE.COM (his
mailbox home server is BIRD04). We use Outlook Express with a IMAP account on
from the client station BIRD08. The inbox is set to synchronize the message.
Reading messages is a two part process :
• Get information on mailbox folders
• Download unread messages
Here are the frames which are generated when getting information on mailbox folders
(IMAP)
1 3.795458 BIRD08 Clt BIRD06 FE X2000 TCP ....S., len: 0, seq:3572614102-3572614102, ack BIRD08 Clt
BIRD06 FE IP
2 3.795458 BIRD06 FE X2000 BIRD08 Clt TCP .A..S., len: 0, seq:2884848472-2884848472, ack BIRD06 FE
BIRD08 Clt IP
3 3.795458 BIRD08 Clt BIRD06 FE X2000 TCP .A...., len: 0, seq:3572614103-3572614103, ack BIRD08 Clt
BIRD06 FE IP
IMAP service on front-end server responds. The IMAP dialog can begin
4 3.795458 BIRD06 FE X2000 BIRD08 Clt TCP .AP..., len: 93, seq:2884848473-2884848566, ack BIRD06 FE
BIRD08 Clt IP
The client requests the listing of capabilities that the front-end server
supports
5 3.795458 BIRD08 Clt BIRD06 FE X2000 TCP .AP..., len: 17, seq:3572614103-3572614120, ack BIRD08 Clt
BIRD06 FE IP
152
The front-end server sends its supported capacities

BIRD08 Clt IP
The client sends its credentials to be authenticated on the front-end server

BIRD06 FE IP
The front-end server initializes the TCP connection with BIRD01 (GC) on port
3268 (LDAP) and sends credentials (frames 8 to 14)
The front-end server sends a LDAP query to BIRD01 (GC)
15 3.805472 BIRD06 FE X2000 BIRD01 GC LDAP ProtocolOp: SearchRequest (3) BIRD06 FE BIRD01 IP
It queries the login “test1” as a “user” attributes

LDAP: Filter Type = And
LDAP: Attribute Type = mailNickName
LDAP: Attribute Value = test1
LDAP: Attribute Type = objectClass
LDAP: Attribute Value = user
The fields requested are:

DistinguishedName, objectGUID, objectClass, modifyTimestamp, legacyExchangeDN,
delivContLength ,Heuristics, delivExtContTypes, DelivEITs, EncapsulationMethod, owner,
ReportToOwner, ReportToOriginator, OOFReplyToOriginator, HideDLMembership,
SecurityProtocol, UnauthOrig, AuthOrig, DLMemRejectPerms, DLMemSubmitPerms,
DeliverAndRedirect, textEncodedORAddress, altRecipient, HomeMTA, HomeMDB,
MapiRecipient, ProxyAddresses, TargetAddress, msExchHomeServerName,
msExchExpansionServerName, objectCategory, deliveryMechanism, autoReply,
MDBStorageQuota, MDBOverQuotaLimit, SubmissionContLength, MDBUseDefaults,
AuthOrigBL, UnauthOrigBL, DLMemRejectPermsBL, DLMemSubmitPermsBL, ProtocolSettings,
deletedItemFlags, GarbageCollPeriod, mDBOverHardQuotaLimit, showInAddressBook,
ntSecurityDescriptor, displayName, objectSID, sn, givenName, mailNickName,
DisplayNamePrintable, adminDisplayName, InternetEncoding, msExchHomePublicMDB,
msExchMasterAccountSid, msExchMailboxSecurityDescriptor, msExchMailboxGuid,
userAccountControl, notes, mail, msExchQueryBaseDN, member, forwardingAddress,
msExchRecipLimit, msExchMemberFilter, msExchMemberBaseDN, managedBy, cn

16 3.815486 BIRD01 GC BIRD06 FE X2000 LDAP ProtocolOp: SearchResponse (4) BIRD01 BIRD06 FE
IP
17 3.815486 BIRD01 GC BIRD06 FE X2000 TCP .A...., len: 1460, seq:1310878666-1310880126, ack BIRD01
BIRD06 FE IP
18 3.815486 BIRD01 GC BIRD06 FE X2000 TCP .AP..., len: 758, seq:1310880126-1310880884, ack BIRD01
BIRD06 FE IP
19 3.815486 BIRD06 FE X2000 BIRD01 GC TCP .A...., len: 0, seq:2884909188-2884909188, ack BIRD06 FE
BIRD01 IP
Front-end server opens a TCP connection with recipient back-end server on

port 143 (IMAP4)
20 3.815486 BIRD06 FE X2000 BIRD04 BE X2000 TCP ....S., len: 0, seq:2884951766-2884951766, ack BIRD06
FE BIRD04 BE X2000 IP
22 3.815486 BIRD06 FE X2000 BIRD04 BE X2000 TCP .A...., len: 0, seq:2884951767-2884951767, ack BIRD06
FE BIRD04 BE X2000 IP
IMAP service on back-end server responds. The IMAP dialog can begin
23 3.815486 BIRD04 BE X2000 BIRD06 FE X2000 TCP .AP..., len: 93, seq:2950776391-2950776484, ack
The front-end server requests the listing of capabilities that the back-end
server supports
24 3.825501 BIRD06 FE X2000 BIRD04 BE X2000 TCP .AP..., len: 14, seq:2884951767-2884951781, ack
The back-end server sends its supported capacities

Front-end forwards the client authentication request to the back-end server

154
Back-end server authenticates the front-end server

Front-end forwards the authentication acknowledge to the client

BIRD08 Clt IP
Client sends some frames to the front-end server to say it is in waiting state,
these frames are forwarded to back-end server and acknowledged through
the front-end server (frames 29 to 36)
Client asks for « Drafts » folder status via the front-end server
BIRD06 FE IP
Back-end server replies to client with requested folder information (frames 39

and 40).
BIRD08 Clt IP
We see clearly in the capture that there is no message (MESSAGES 0), and no unread
messages (UNSEEN 0) in that folder
Client asks for « Inbox » folder status via the front-end server (frames 49 and
50)
BIRD06 FE IP

and 52).
BIRD08 Clt IP
We see in the capture that there is one message (MESSAGES 1) which is unread
(UNSEEN 1) in that folder.
Client asks for « Sent Items » folder status via the front-end server (frames 61
and 62)
BIRD06 FE IP

and 64).
BIRD08 Clt IP
156
We see in the capture that there are 30 read message (MESSAGES 30 UNSEEN 0) in
that folder.
Frames 69 to 75 are closing connections between Client and FE, and FE and
BE
Now, the first part of client refresh is over. IMAP folders are up to date.
The next part is the downloading of new messages. It is not described because its logical
path is very similar to the first part:
• Opens connection,
• Asks capacities
• Authentication
• Downloads messages
• Disconnection
This phase operates also like above : all client requests received by the front server are
forwarded to its back-end server and vice-versa. The front-end server acts as router also
with IMAP protocol.
front-end and back-end server.
Figure 13 : FE/BE Message Reading Traffic Frames Overview
1 3 5 7 8 ... 14 29 ... 36 37 41 ... 48 49 53 ... 60 61 65 ... 68
BIRD06
BIRD08 2 4 6 8 ... 14 28 29 ... 36 40 41 ... 48 52 53 ... 60 64 65 ... 68 Front-End
Client
20 21
22 23
15
24 25
19 16
17 26 27
18
38 39
50 51
62 63
BIRD01
GC 69 69
... ...
75 75
BIRD04
Back-End
14.2.1. Analysis
Operation mode with IMAP is closed from the one used with SMTP requests. The main
difference is that IMAP requests are immediately forwarded to the back-end server or to
the client via the front-end server. An IMAP client only communicates with the front-
end server. The front-end server send LDAP queries to GC to know where is the login
user mailbox home server. It acts as a IMAP router.
158
14.3. Global analysis about Front-End / Back-End

communications
Any clients using an Internet protocol (POP3, IMAP4, SMTP) can take advantages of a
front-end server.
MAPI clients cannot use these configuration ; they always contact their real mailbox
server.
A front-end server is designed as a router for internet protocols : it centralize client
access, forwarding their requests to the right back-end server. When a message is sent to
recipients who are not on the same Exchange box, the front-end sends the message to
each concerned mailbox home server. When a client makes a LDAP query to the front-
end server, the front-end server forward the query to the closest GC then forward reply
from GC to client.
As a front-end server acts as a protocol router, the whole traffic generated by a client is
forwarded to the back-end side (and vice-versa). There is no compression, neither
modifications.
You can create mailboxes on front-end servers. But, those mailboxes will only be
reachable with MAPI clients. Internet client cannot access to mailboxes located on a
front-end server.
15. Routing Group Connector
This part must provide us traffic analysis when sending messages through a routing
group connector. First of all, we will examine it in a front-end / back-end situation.
15.1. Lab configuration
In order to study routing group connector (RG), we create 2 routing groups :

• Paris, which includes the servers BIRD04, BIRD05 and BIRD06.
• Nantes, which includes the server BIRD07.
A Routing Group Connector is configured to link the two routing groups Paris and
Nantes. To make easier captures, we define bridgehead servers:
• On the PARIS Routing Group, the server BIRD04 is set to be the only server to
be able to send messages to the NANTES Routing Group.
The remote bridgehead server in NANTES Routing Group is BIRD07.
Figure 14 : Paris Routing Group Configuration (1/2)

160
Figure 15 : Paris Routing Group Configuration (2/2)
• On the NANTES Routing Group, the server BIRD07 is set to be the only
server to be able to send messages to the NANTES Routing Group.
The remote bridgehead server in PARIS Routing Group is BIRD04.
Figure 16 : Nantes Routing Group Configuration (1/2)

Figure 17 : Nantes Routing Group Configuration (2/2)
15.2. Message Sending
To capture the traffic in a front-end/back-end server environment where Routing Group

Connectors are implicated, we send a message from TEST1@EUROPE.COM (his
mailbox home server is BIRD06, Routing Group PARIS) to the recipient
TEST7@EUROPE.COM (his mailbox home server is BIRD07, Routing Group
NANTES). We use Outlook Express with a IMAP account on from the client station
BIRD08. The account is set to not copy sent items in « Sent items » folder.
We comment below the network traffic which is generated by sending this message.
Note : some of frames do not appear in the numeric order because they were captured in
the same time. To comment accurately, we replace the frames in the real logical order.
(SMTP)
1 1.361958 BIRD08 Clt BIRD06 FE X2000 TCP ....S., len: 0, seq:1755925180-1755925180, ack BIRD08
Clt BIRD06 FE IP
2 1.361958 BIRD06 FE X2000 BIRD08 Clt TCP .A..S., len: 0, seq:1012320851-1012320851, ack
3 1.361958 BIRD08 Clt BIRD06 FE X2000 TCP .A...., len: 0, seq:1755925181-1755925181, ack BIRD08
Clt BIRD06 FE IP
The SMTP service on the front-end server responds. The SMTP dialog can
begin
4 1.361958 BIRD06 FE X2000 BIRD08 Clt SMTP Rsp: Service ready, 116 bytes BIRD06 FE
BIRD08 Clt IP
162
The client presents itself to the front-end server

5 1.361958 BIRD08 Clt BIRD06 FE X2000 SMTP Cmd: Hello, host identifier, 13 bytes BIRD08 Clt
BIRD06 FE IP
The sender name is sent to the front-end server

7 1.371973 BIRD08 Clt BIRD06 FE X2000 SMTP Cmd: Mail from <test1@europe.com>, 31 bytes
The recipient name is sent to the front-end server

9 1.371973 BIRD08 Clt BIRD06 FE X2000 SMTP Cmd: Recipient <test7@europe.com>, 29 bytes
The message body is sent to the front-end server

11 1.371973 BIRD08 Clt BIRD06 FE X2000 SMTP Cmd: Data: Mail data to follow, 6 bytes BIRD08
Clt BIRD06 FE IP
12 1.371973 BIRD06 FE X2000 BIRD08 Clt SMTP Rsp: Enter mail ..., 46 bytes BIRD06 FE
BIRD08 Clt IP
13 1.381987 BIRD08 Clt BIRD06 FE X2000 SMTP Data - continued from frame 11, 1155 bytes
15 1.582275 BIRD08 Clt BIRD06 FE X2000 SMTP Data - continued from frame 13, 6 bytes BIRD08
Clt BIRD06 FE IP
The client quits the connection to the front-end server

17 1.592290 BIRD08 Clt BIRD06 FE X2000 SMTP Cmd: Quit, losing connection, 6 bytes BIRD08
Clt BIRD06 FE IP
The front-end server acknowledges SMTP connection closing

18 1.592290 BIRD06 FE X2000 BIRD08 Clt SMTP Rsp: Service closing transmission channel, 66 byt
The TCP connection is closed between the client and the front-end server
19 1.592290 BIRD06 FE X2000 BIRD08 Clt TCP .A...F, len: 0, seq:1012321271-1012321271, ack
20 1.592290 BIRD08 Clt BIRD06 FE X2000 TCP .A...., len: 0, seq:1755926426-1755926426, ack
21 1.592290 BIRD08 Clt BIRD06 FE X2000 TCP .A...F, len: 0, seq:1755926426-1755926426, ack
The front-end server sends a LDAP query to BIRD01 (GC) on port 3268
(specific GC port)
23 1.592290 BIRD06 FE X2000 BIRD01 GC LDAP ProtocolOp: SearchRequest (3) BIRD06 FE
BIRD01 IP
+ Frame: Base frame properties
+ ETHERNET: ETYPE = 0x0800 : Protocol = IP: DOD Internet Protocol
+ IP: ID = 0x858E; Proto = TCP; Len: 1326
+ TCP: .AP..., len: 1286, seq: 412539185-412540471, ack:3661302845, win:17520, src:21448
dst: 3268
+ LDAP: ProtocolOp: SearchRequest (3)
It searches sender (test1@europe.com) and recipient

(test7@europe.com)
LDAP: Filter Type = Or
The requested fields are:
DistinguishedName, forwardingAddress, objectClass, mail, member, ProxyAddresses,
textEncodedORAddress, legacyExchangeDN, objectGUID, altRecipient, DeliverAndRedirect,
HideDLMembership, delivContLength, SubmissionContLength, TargetAddress, HomeMDB,
msExchHomeServerName, AuthOrig, UnauthOrig, DLMemSubmitPerms, DLMemRejectPerms,
msExchRecipLimit, msExchMemberFilter, msExchMemberBaseDN, ReportToOriginator,
ReportToOwner, managedBy, cn, MapiRecipient, InternetEncoding, HomeMTA,
msExchExpansionServerName, msExchMailboxGuid, OOFReplyToOriginator, modifyTimestamp,
Heuristics, delivExtContTypes, DelivEITs, EncapsulationMethod, owner, SecurityProtocol,
objectCategory, deliveryMechanism, autoReply, MDBStorageQuota, MDBOverQuotaLimit,
MDBUseDefaults, AuthOrigBL, UnauthOrigBL, DLMemRejectPermsBL, DLMemSubmitPermsBL,
ProtocolSettings, deletedItemFlags, GarbageCollPeriod, mDBOverHardQuotaLimit,
showInAddressBook, ntSecurityDescriptor, displayName, objectSID, sn, givenName,
mailNickName, DisplayNamePrintable, adminDisplayName, msExchHomePublicMDB,
msExchMasterAccountSid, msExchMailboxSecurityDescriptor, userAccountControl, notes,
msExchQueryBaseDN

24 1.602304 BIRD01 GC BIRD06 FE X2000 LDAP ProtocolOp: SearchResponse (4) BIRD01
BIRD06 FE IP
25 1.602304 BIRD01 GC BIRD06 FE X2000 TCP .A...., len: 1460, seq:3661304305-3661305765, ack
BIRD01 BIRD06 FE IP
BIRD01 BIRD06 FE IP
BIRD01 BIRD06 FE IP
BIRD01 BIRD06 FE IP
29 1.602304 BIRD01 GC BIRD06 FE X2000 TCP .AP..., len: 33, seq:3661310145-3661310178, ack
BIRD01 BIRD06 FE IP
30 1.602304 BIRD06 FE X2000 BIRD01 GC TCP .A...., len: 0, seq: 412540471-412540471, ack:
BIRD06 FE BIRD01 IP
31 1.602304 BIRD06 FE X2000 BIRD01 GC TCP .A...., len: 0, seq: 412540471-412540471, ack:
BIRD06 FE BIRD01 IP
Front-end server opens a TCP connection with local bridgehead server of

front-end routing group (BIRD04) on port 25 (SMTP)
32 3.655256 BIRD06 FE X2000 BIRD04 BE X2000 TCP ....S., len: 0, seq:1013041778-1013041778, ack
SMTP service on local bridgehead server responds. The SMTP dialog can
begin
35 3.655256 BIRD04 BE X2000 BIRD06 FE X2000 SMTP Rsp: Service ready, 116 bytes BIRD04 BE
X2000 BIRD06 FE IP
Local bridgehead server server authenticates front-end server

36 3.655256 BIRD06 FE X2000 BIRD04 BE X2000 SMTP Data, 24 bytes BIRD06 FE BIRD04 BE
X2000 IP
37 3.655256 BIRD04 BE X2000 BIRD06 FE X2000 SMTP Rsp: Requested mail action okay, completed,
231 b BIRD04 BE X2000 BIRD06 FE IP
164
39 3.655256 BIRD04 BE X2000 BIRD06 FE X2000 SMTP Rsp: Unknown response type, 68 bytes
The sender name is sent to the local bridgehead server

40 3.655256 BIRD06 FE X2000 BIRD04 BE X2000 SMTP Cmd: Mail from <test1@europe.com>, 40 bytes
The recipient name is sent to the local bridgehead server

42 3.655256 BIRD06 FE X2000 BIRD04 BE X2000 SMTP Cmd: Recipient <test7@europe.com>, 28 bytes
The message body is sent to the local bridgehead server

45 3.665270 BIRD04 BE X2000 BIRD06 FE X2000 SMTP Rsp: Enter mail ..., 22 bytes BIRD04 BE
X2000 BIRD06 FE IP
Front-end server quits the connection to the local bridgehead server

52 3.815486 BIRD06 FE X2000 BIRD04 BE X2000 SMTP Cmd: Quit, losing connection, 6 bytes
Local bridgehead server acknowledges SMTP connection closing

53 3.815486 BIRD04 BE X2000 BIRD06 FE X2000 SMTP Rsp: Service closing transmission channel, 66
byt BIRD04 BE X2000 BIRD06 FE IP
The TCP connection is closed between the front-end and the back-end server
54 3.815486 BIRD06 FE X2000 BIRD04 BE X2000 TCP .A...F, len: 0, seq:1013044439-1013044439, ack
55 3.815486 BIRD04 BE X2000 BIRD06 FE X2000 TCP .A...F, len: 0, seq:3887739882-3887739882, ack
At this point the traffic is 100% equal to the traffic in above section (with recipient in
same routing group). We see that front-server contacts the local bridgehead server to
send a message to a recipient in a remote routing group.
The next part is to transfer the message from local bridgehead server (in RG Paris) to the
remote bridgehead server (in RG Nantes). It means from BIRD04 to BIRD07.
Local bridgehead server (BIRD04, in RG Paris) initializes the TCP connection

with the remote bridgehead server (BIRD07, in RG Nantes) front-end server
on port 25 (SMTP)
59 5.808352 BIRD04 BE X2000 BIRD07 X2000 TCP ....S., len: 0, seq:3888348874-3888348874, ack
BIRD04 BE X2000 BIRD07 BE IP
60 5.808352 BIRD07 X2000 BIRD04 BE X2000 TCP .A..S., len: 0, seq:2373514979-2373514979, ack
BIRD07 BE BIRD04 BE X2000 IP
61 5.808352 BIRD04 BE X2000 BIRD07 X2000 TCP .A...., len: 0, seq:3888348875-3888348875, ack
The SMTP service on the remote bridgehead server responds. The SMTP
dialog can begin
62 5.808352 BIRD07 X2000 BIRD04 BE X2000 SMTP Rsp: Service ready, 116 bytes BIRD07 BE
BIRD04 BE X2000 IP
Local bridgehead server (BIRD04, in RG Paris) asks for supported Extended

SMTP commands and authentication methods to BIRD07, then sends its
identity
63 5.818366 BIRD04 BE X2000 BIRD07 X2000 SMTP Data, 24 bytes BIRD04 BE X2000 BIRD07 BE
IP
65 5.838395 BIRD04 BE X2000 BIRD07 X2000 SMTP Data - continued from frame 63, 14 bytes
BIRD07 replies with its supported commands and authenticates BIRD04

64 5.818366 BIRD07 X2000 BIRD04 BE X2000 SMTP Rsp: Requested mail action okay, completed, 275
b BIRD07 BE BIRD04 BE X2000 IP
66 5.838395 BIRD07 X2000 BIRD04 BE X2000 SMTP Rsp: Unknown response type, 22 bytes
69 5.838395 BIRD07 X2000 BIRD04 BE X2000 TCP .A...., len: 0, seq:2373515393-2373515393, ack
The next frame is the authentication validation frame :
The sender name is sent to BIRD07

75 5.848410 BIRD04 BE X2000 BIRD07 X2000 SMTP Cmd: Mail from <test1@europe.com>, 40 bytes
by BIRD07 BE BIRD04 BE X2000 IP
166
The recipient name then the message body are sent to BIRD07
77 5.858424 BIRD04 BE X2000 BIRD07 X2000 SMTP Cmd: Recipient <test7@europe.com>, 28 bytes
80 5.858424 BIRD07 X2000 BIRD04 BE X2000 SMTP Rsp: Enter mail ..., 22 bytes BIRD07 BE
BIRD04 BE X2000 IP
BIRD07 acknowledges that the message is well received and is now queued
Local birdgehead server (BIRD04) closes the connection to remote

bridgehead server (BIRD07)
88 5.928525 BIRD04 BE X2000 BIRD07 X2000 SMTP Cmd: Quit, losing connection, 6 bytes BIRD04
BE X2000 BIRD07 BE IP
BIRD07 acknowledges SMTP connection closing

89 5.938539 BIRD07 X2000 BIRD04 BE X2000 SMTP Rsp: Service closing transmission channel, 66 byt
The TCP connection is closed between the two bridgehead servers

90 5.938539 BIRD04 BE X2000 BIRD07 X2000 TCP .A...F, len: 0, seq:3888353472-3888353472, ack
92 5.938539 BIRD07 X2000 BIRD04 BE X2000 TCP .A...F, len: 0, seq:2373516168-2373516168, ack
93 5.938539 BIRD04 BE X2000 BIRD07 X2000 TCP .A...., len: 0, seq:3888353473-3888353473, ack
168
front-end and bridgehead servers.
Figure 18 :Routing Group Message Sending Traffic Frames Overview
BIRD08 1
Client Depends on message
3 body length
5
2
7
4
9
6
11
8 ...
10 16
11 17
... 20
16
21
18
19
22
BIRD06
Front-End
23
30 24 32 33
31 25 34 35
26
36 37
27
28 38 39
29
40 41
BIRD01
GC
42 43
44 45
...
45
Depends on message 51
body length ...
51 53
52 55
54 56
57 58
92 91 89 87 86 82 80 78 76 74 72 70 69 66 64 62 60
BIRD07 93 90 88 85 84 83 81 79 77 75 73 71 68 67 65 63 61 59 BIRD04
Back-End Back-End
15.2.1. Analysis
As designed, the client does communicate only with the front-end server.
The front-end server send LDAP queries to GC to know who are the sender and
recipient, and especially to know where is the home server of the recipient. It sees that
the recipient is on a different routing group. So it needs to contact the local bridgehead in
its own routing group (the front-end routing group) which knows how to contact the
remote routing group. Then, the front-end server send the message to the local
bridgehead server. Then the local bridgehead server sends all the messages in its queue
to the remote bridgehead server. As seen above, there is no data compression between
servers through a routing group: it is pure SMTP dialog.
In that test, the front-end server cannot send messages to the remote routing group. If we
have enabled it, the traffic would have been lesser. The front-end could send directly the
message to the server in the remote routing group. Then, we save one exchange
(between front-end sever and local bridgehead server).
170
16. Appendix
This section includes detailed information about DSProxy and client access to
Active Directory.
16.1. DSProxy
At startup, the Exchange System Attendant finds the most appropriate Active
Directory server in the domain through the Domain Name System (DNS), then
resolves and passes its name through to the DSProxy process (Dsproxy.dll). This is
signaled by event 9010 for the MSExchangeSA process.
It is also possible to ascertain which Active Directory domain controller a
particular Exchange server is using by means of Exchange computer properties in
the Microsoft Exchange System Manager snap-in.
In some situations, the administrator can specifically set the server that DSProxy
uses. You can accomplish this by changing the following registry entries:
To specify the global catalog server for earlier MAPI clients:
1. On the Run line, type regedit.exe, and then click OK.
2. In the registry editor, navigate to HKEY_LOCAL_MACHINE\SYSTEM
\CurrentControlSet\Services\MSExchangeSA\Parameters.
3. Select the NSPITargetServer entry.
4. Assign the name of the global catalog server to specify the global catalog
server for earlier MAPI clients that use only NSPI.
5. Close the registry editor.
Caution Do not use a registry editor to edit the registry directly unless you
have no alternative. The registry editors bypass the standard safeguards
provided by administrative tools. These safeguards prevent you from entering
conflicting settings or those likely to degrade performance or damage your
system. Editing the registry directly can have serious, unexpected
consequences that prevent the system from starting, thus requiring a re-install
of Windows 2000 or Exchange 2000. To configure or customize
Windows 2000, use the programs in Control Panel or Microsoft Management
Console (MMC) whenever possible.
To specify the global catalog server for recent MAPI clients

(Outlook 2000) which can use RFR:
3. Select the RFRTargetServer entry.
4. Assign the name of the global catalog server to specify the global catalog
server for recent MAPI clients (Outlook 2000) that use RFR.
The DSProxy NSPI works by blindly forwarding MAPI directory system (DS)
requests to a global catalog server. This means that the remote procedure call
(RPC) packet is not opened or evaluated because this would incur a significant
overhead on the Exchange 2000 server, and complicate the security structure. The
process begins by creating a listening thread for each supported network protocol,
and a single working thread for each processor. This can accommodate up to 512
client connections and dynamically adds more threads as required. A socket-
mapping table keeps a reference of connections between clients and servers,
ensuring that the correct responses from Active Directory pass to the associated
client.
DSProxy works over TCP/IP, Internetwork Packet Exchange (IPX), and AppleTalk
protocols; however, it does not work over network basic input/output system
(NetBIOS).
The following sections explain how different mail clients access information stored
in Active Directory.
16.2. Client Access to Active Directory
16.2.1. Outlook 98 and Outlook 2000

Newer versions of the Outlook client, such as Outlook 98 and Outlook 2000
(version 8.5.6204.0 or later), use a slightly different method for address book
access. Initially, Outlook expects to find the directory service on the home
Exchange server. Because the version of the Exchange server running on the
system can be determined only after loading Emsmdb32.dll (which is after the
address book provider—Emsabp.dll), Outlook will go through the DSProxy
process for the first session. Once the client has contacted the DSProxy service (it
will try all available transport protocols), a special referral passes back to the client,
informing that client that all future directory requests should be sent to the global
catalog server. Outlook stores this referral in the MAPI profile in the registry.
172
The referral mechanism reduces the load on the Exchange 2000 server and address
book lookup latency; however, when an explicit server name is entered into the
profile, Outlook requires a restart if that Active Directory server fails. If that
occurs, the Exchange 2000 server passes Outlook a new referral.
Figure 19 : DSAccess Overview
Exchange 2000 Global Catalog

Server
Messages Directory
requests
Outlook
Some scenarios require Outlook clients, even the latest versions, to go through the
DSProxy process without being referred. For example, when a firewall exists
between client computers and Active Directory servers, the firewall can be opened
up to allow the Exchange 2000 server to access Active Directory.
To prevent Exchange from returning referrals:
3. Click Edit, point to New, and choose DWORD Value.
4. Type NoRFRService to name the entry.
5. Right-click NoRFRService, click Modify, and then assign a value of 1 to
prevent Exchange from returning directory referrals.
Caution Do not use a registry editor to edit the registry directly unless you
have no alternative. The registry editors bypass the standard safeguards
provided by administrative tools. These safeguards prevent you from entering
conflicting settings or those likely to degrade performance or damage your
system. Editing the registry directly can have serious, unexpected
consequences that prevent the system from starting, thus requiring a re-install
of Windows 2000 or Exchange 2000. To configure or customize
Windows 2000, use the programs in Control Panel or Microsoft Management

Console (MMC) whenever possible.
16.2.1.1. First discovery of the GC the Outlook 2000 client
Frame 1 :
The capture starts with a BONE frame; this is an administrative frame that
Network Monitor puts onto the network wire for its own use. A BONE frame,
which uses a target MAC address of Multicast 2 or 030000000002 on Ethernet,
will not cross a router unless the router is configured to bridge BONE frames.
Note : Just if you need more info on the bone frame:
Network Monitor uses the BONE frame to:
• Find out who is using Network Monitor on the local network.
• Implement security (in Network Monitor version 2.0 only).
Frame 2 & 3 :
Client goes for IP resolution via ARP (usual name resolution)
Frame 4 – 6 :
This is the classical TCP 3way handshake with 2 network devices using this
protocol.
Frame 7 – 10 :
RPC bind to UUID = E1AF8308-5D1F-11C9-91A4-08002B14A0F1 (End
Point Mapper) on port 135 (this is the port ALWAYS used by EPM to listen to
RPC calls). Any client requesting an RPC call will first connect to this port
then pass to the server the requested service. The answer will be the port of the
service requested.
Here client gets a "bind ack" from the server and then request the opnum 0x3.
The source code indicate the called function is "ept_map" (the port on which
the service the client wants to connect to is listening).
Remember RPC calls are encrypted so we cannot know what info is passed. We
can only see header infos.
Frame 11 – 14 :
This is the TCP usual end of conversation. Acknowledge the previous frame
and send a end signal (No more data from sender).
174
Frame 15 – 17 :
TCP 3 way handshake between client and Exchange Server.
Frame 18 :
Client bind to UUID = 1544F5E0-613C-11D1-93DF-00C04FD7BD09. This is
basically the only UUID of the Referral Interface (RFR).
Bind ack from server
Frame 20 :
Client call function rfrGetNewDSA (the client request a name for a directory)
=> the return is the GC FQDN.
From now, we will only focus on RPC frames as they are the most interesting.
Frame 31 – 34 :
After ARP detection and TCP 3way handshake, client call the server returned
by Exchange 2000 as the directory server on the EMP (UUID = E1AF8308-
5D1F-11C9-91A4-08002B14A0F1) basically to know on which port the client
can contact the NSPI interface. The NSPI interface is in charge of accepting an
authenticating client connections.
The answer of the GC is a bind ack.
We have then TCP traffic of disconnection & 3way handshake
Frame 42 & 43 :
Client send a call to the directory on the port returned by the EPM using the
UUID = F5CC5A18-4264-101A-8C59-08002B2F8426. The result is a bind
ack. This does not mean, the client is connected to the directory ! This only
means the directory is ready to accept calls from Outlook clients.
Be aware the RPC call in frame 42 is NOT authenticated ! This is a regular
procedure. Clients always connect to the directory anonymously the first time.
Frame 44 & 45 :
Client tries to bind to the directory. Remember it has not authenticated on the
RPC_bind call.
The result of the NSPI_Bind is 0x80040111 = non zero = fails ! We could make
some anonymous RPC calls on the directory but definitely not a bind.
Frame 49 & 50 :
Outlook then goes for an authenticated bind to the Exchange Directory. Note
we don't go back to the EPM as we know which port the service listen to.
Frame 52 & 54 :
The client binds successfully to the directory ! The result of the NSPIBind is
equal to zero. Remember this time, RPC bind was authenticated.
At this point, client is authenticated. In Exchange, the directory is always in charge

of authentication. The store does not check for authentication. The client can then
process to connect to the store.
Frame 58 – 61 :
After a TCP 3way handshake the client connect to the store EPM to get the
store access point (port number the service is listening to).
Frame 69 :
The client ended the call with the server and started again a TCP connection.
The RPC call then goes to the UUID = A4F1DB00-CA47-1067-B31F-
00DD010662DA (The store UUID). Note the RPC call is however
authenticated :)
Frame 70 :
Server acknowledge the client RPC request. The client is bound to the store.
We can now proceed with usual logon sequence.
Frame 72 :
EcDoConnect: Client goes for a MAPI connection to the store and passes the
DN of the Mailbox it wants to open .
Frame 73 :
Exchange 2000 server passes a request to a DC. This confirm that the
Exchange server is passing the authentication provided by the client to a DC to
validate the EcDoConnect request to this specific mailbox.
We have the answer of the DC in frame 75 and then Exchange reply to the
client on Frame 76 with the DN of the mailbox to open.
176
Frame 77 :
The RPC call from the client is EcDoRPC which is a package of several MAPI
Calls to improve store access efficiency. One of the calls is EcDoLogon. This
would then explain the further traces.
Exchange 2000 fires an LDAP request to the GC sending the
msExchMasterAccountSid and requesting data like objectDistinguishedName
or ObjectGUID.
As soon as we recieve the anwser from the GC (LDAP) we give the answer to
the client in RPC mode (Frame 82).
Frame 83 - 86:
We only can say this is a EcDoRPC call. Nothing can be dumped out of the
crypted data ...
As we don't see any access to a DC or a GC, this is some calls like
"EcOpenFolder" or "EcQueryRows".
Frame 87:
EcRegisterPushNotification: The client register to the server a port on which he
can receive new mail notification.
Frame 89 - 100:
We get consecutive EcDoRPC calls. We are not really able to see what's done
at this point of the logon sequence : an EMSMDB32 trace would be more
efficient.
Frame 101 & 102:

We have NspiGetHierarchyInfo RPC calls. The client connects to what he
thinks is a DS to get the hierarchy.
Table 7 : Outlook 2000 logon Frames
Frame Time Source Destination Protocol Description

1 0.593750 Bone Security Check (0x03)
2 3.609375 ARP_RARP ARP: Request, Target IP: 10.10.10.43
3 3.609375 ARP_RARP ARP: Reply, Target IP: 10.10.10.19 Target Hdwr Ad
4 3.609375 CLIENT EX2000 TCP ....S., len: 0, seq:3978097722-3978097722, ack
5 3.609375 EX2000 CLIENT TCP .A..S., len: 0, seq:3280470153-3280470153, ack
6 3.609375 CLIENT EX2000 TCP .A...., len: 0, seq:3978097723-3978097723, ack
7 3.609375 CLIENT EX2000 MSRPC c/o RPC Bind: UUID E1AF8308-5D1F-11C9-91A
8 3.609375 EX2000 CLIENT MSRPC c/o RPC Bind Ack: call 0x1 assoc grp 0x9531
9 3.609375 CLIENT EX2000 MSRPC c/o RPC Request: call 0x1 opnum 0x3 contex
10 3.609375 EX2000 CLIENT MSRPC c/o RPC Response: call 0x1 context 0x0 hint
11 3.609375 CLIENT EX2000 TCP .A...F, len: 0, seq:3978097951-3978097951, ack
12 3.609375 EX2000 CLIENT TCP .A...., len: 0, seq:3280470366-3280470366, ack
13 3.609375 EX2000 CLIENT TCP .A...F, len: 0, seq:3280470366-3280470366, ack
18 3.609375 CLIENT EX2000 MSRPC c/o RPC Bind: UUID 1544F5E0-613C-11D1-93D
19 3.609375 EX2000 CLIENT MSRPC c/o RPC Bind Ack: call 0x1 assoc grp 0x27CEF

26 3.625000 ARP_RARP ARP: Request, Target IP: 10.10.10.26
27 3.625000 ARP_RARP ARP: Reply, Target IP: 10.10.10.19 Target Hdwr Ad
28 3.625000 CLIENT GC TCP ....S., len: 0, seq:3978208196-3978208196, ack
29 3.625000 GC CLIENT TCP .A..S., len: 0, seq:3395322137-3395322137, ack
30 3.625000 CLIENT GC TCP .A...., len: 0, seq:3978208197-3978208197, ack
31 3.625000 CLIENT GC MSRPC c/o RPC Bind: UUID E1AF8308-5D1F-11C9-91A
32 3.625000 GC CLIENT MSRPC c/o RPC Bind Ack: call 0x1 assoc grp 0x9094
33 3.625000 CLIENT GC MSRPC c/o RPC Request: call 0x1 opnum 0x3 contex
34 3.625000 GC CLIENT MSRPC c/o RPC Response: call 0x1 context 0x0 hint
35 3.625000 CLIENT GC TCP .A...F, len: 0, seq:3978208425-3978208425, ack
36 3.625000 GC CLIENT TCP .A...., len: 0, seq:3395322350-3395322350, ack
37 3.625000 GC CLIENT TCP .A...F, len: 0, seq:3395322350-3395322350, ack
42 3.625000 CLIENT GC MSRPC c/o RPC Bind: UUID F5CC5A18-4264-101A-8C5
43 3.625000 GC CLIENT MSRPC c/o RPC Bind Ack: call 0x1 assoc grp 0x848A
44 3.625000 CLIENT GC R_NSPI RPC Client call nspi:NspiBind(..)
45 3.625000 GC CLIENT R_NSPI RPC Server response nspi:NspiBind(..)
49 3.640625 CLIENT GC MSRPC c/o RPC Bind: UUID F5CC5A18-4264-101A-8C5
50 3.640625 GC CLIENT MSRPC c/o RPC Bind Ack: call 0x2 assoc grp 0x848A
51 3.640625 CLIENT GC TCP .AP..., len: 188, seq:3978315939-3978316127, ack
52 3.640625 CLIENT GC R_NSPI RPC Client call nspi:NspiBind(..)

53 3.640625 GC CLIENT TCP .A...., len: 0, seq:3395441419-3395441419, ack
54 3.640625 GC CLIENT R_NSPI RPC Server response nspi:NspiBind(..)
58 3.703125 CLIENT EX2000 MSRPC c/o RPC Bind: UUID E1AF8308-5D1F-11C9-91A
59 3.703125 EX2000 CLIENT MSRPC c/o RPC Bind Ack: call 0x1 assoc grp 0x9532
69 3.703125 CLIENT EX2000 MSRPC c/o RPC Bind: UUID A4F1DB00-CA47-1067-B31
70 3.718750 EX2000 CLIENT MSRPC c/o RPC Bind Ack: call 0x1 assoc grp 0x2805C
71 3.718750 CLIENT EX2000 TCP .AP..., len: 188, seq:3978436935-3978437123, ack
73 3.718750 EX2000 DC MSRPC c/o RPC Request: call 0x9 opnum 0x27 conte
75 3.718750 DC EX2000 MSRPC c/o RPC Response: call 0x9 context 0x0 hint
78 3.718750 EX2000 GC LDAP ProtocolOp: SearchRequest (3)
79 3.734375 GC EX2000 LDAP ProtocolOp: SearchResponse (4)

80 3.734375 GC EX2000 TCP .AP..., len: 764, seq:3230887744-3230888508, ack
81 3.734375 EX2000 GC TCP .A...., len: 0, seq:3115794305-3115794305, ack
99 3.906250 EX2000 DC TCP .A...., len: 0, seq:3205045752-3205045752, ack
101 9.859375 CLIENT GC R_NSPI RPC Client call nspi:NspiGetHierarchyInfo(..)
102 9.859375 GC CLIENT R_NSPI RPC Server response nspi:NspiGetHierarchyInfo(..)
103 9.859375 GC CLIENT TCP .AP..., len: 204, seq:3395442975-3395443179, ack
105 9.875000 CLIENT EX2000 MSRPC c/o RPC Request: call 0xA opnum 0x2 contex
106 9.875000 EX2000 CLIENT MSRPC c/o RPC Response: call 0xA context 0x0 hint

108 10.593750 Bone Security Check (0x03)
109 11.187500 EX2000 DC ICMP Echo: From 10.10.10.43 To 10.10.10.20
110 11.187500 DC EX2000 ICMP Echo Reply: To 10.10.10.43 From 10.10.10.20
Figure 20 : First discovery of the GC by Outlook 2000 client
1- Client Connects to the RFR Interface

2 - E xc h a n g e S e r v e r r e t u r n s
name of Global Catalog
7- Client access to the Store
8- Server ack RPC call
9- MAPI logon
12- Validate MAPI logon call
Workstation using 13- Open Message Store Exchange 2000 server

Outlook 2000
16- Accept Client access
11 10
- -
V A
og
al ut M
id he AP
tal
at
e nt I l
Ca
us ica ogo
l
er tio n
og
ba
6 -
ac n
tal
4-
co
A
G lo
un
cc
Ca
Ac
5-
3-
es
the
al
ce
s
s
b
i
s
in
e
n
ut
n
Glo
re
rty
Cl to t
fu
ho
m
Cl to th
h
pe
e
e
se
ris
fro
d
e
n
po
t c Gl
t c Gl
d
e
c
er
by
ct
on ob
by
on oba
je
sw
ne al
th
t
e
Domain Controler
ob
An
he
ct l Ca
he
t a Cat
-
Gl
y
n
ve
Gl
an tal
ob
15
ut alo
ie
mo
tic
ob
on og
g
al
at
us
al
etr
ly
ed
at
- R
Ca
ta
alo
14
g
log
Global Catalog
16.2.1.2. Other sessions opened with the same user profile with
Outlook 2000
When you reopen Outlook 2000, the client haven’t to ask a GC to the
Exchange 2000 server. Outlook will use the GC name store in the registry key.
The traffic is exactly the same than for the first session but start at frame 31.
16.2.2. Exchange Client and Outlook 97 and

Outlook 98
Older MAPI clients, such as Exchange 97, Exchange 98, Outlook 97, and Outlook
98 (version 8.5.6204.0 or earlier), make MAPI Directory Service (MAPI DS)
requests to an Exchange server. These requests range from resolving a fragment of
text typed on the To line of a message into a user name, to showing objects from
the global address list. To make Exchange 2000 compatible with the existing MAPI
client base, an Exchange 2000 server sends any MAPI DS requests through to a
local global catalog server on the network. The DSProxy process on the
Exchange 2000 server accomplishes this task and forwards the packet; it does not
change the request into Lightweight Directory Access Protocol (LDAP). Active
Directory supports a number of protocols, including LDAP and MAPI DS, so an
Outlook directory request is valid, even against an Active Directory server.
The global catalog server returns the result to the Exchange 2000 server, and then
sends the result on to the MAPI client. This process is seamless to the user and
takes very little time to complete.
16.2.3. Traffic and Load Generated Through the

DSProxy Process
The following steps describe the communication process for a one-recipient name
lookup:
1. The MAPI client sends one network packet to the Exchange 2000 server.
The packet contains the name to look up, in plain text.
2. The Exchange 2000 server sends the request to a local global catalog
server.
3. The local global catalog server returns the result to the Exchange 2000
server.
4. The Exchange 2000 server returns the result to the MAPI client.
5. The MAPI client returns an acknowledgement to the Exchange 2000 server.
6. The Exchange 2000 server sends the acknowledgement to the local global
catalog server.
184
Figure 21 : DSProxy Access
Global Catalog
2
3
6
1
4
5 Exchange 2000
Server
Outlook
17. Index
17.1. Graphs
Graph 1 : Logon (comparative: Outlook 2000, Outlook 97)...................................25

Graph 2 : Address Resolution (comparative: Outlook 2000, Outlook 97)..............40
Graph 3: Ambiguous Name Resolution (comparative: Outlook 2000, Outlook 97)
........................................................................................................................40
Graph 4 : Address Lookup, Address Book View Lookup, Address Details
(comparative: Outlook 2000, Outlook 97) ......................................................41
Graph 5 : Generic Mail Item / Outlook 2000 - Send item without attachment
(comparative: RTF, HTML, Plain Text) .........................................................53
Graph 6 : Generic Mail Item / Outlook 2000 - Send a 1K item with attachment
(comparative: RTF; HTML, Plain Text) .........................................................54
Graph 7 : Generic Mail Item / Outlook 2000 - Read item without attachment
(comparative: RTF, HTML, Plain Text) .........................................................55
Graph 8: Generic Mail Item / Outlook 2000 - Open and Read a 1K message with
an attachment (comparative: RTF, HTML, Plain Text)..................................55
Graph 9 : Generic Mail Item / Outlook Express / IMAP - Send item without
attachment (comparative: HTML, Plain Text)................................................60
Graph 10 : Generic Mail Item / Outlook Express / IMAP - Send item with
Graph 11 : Generic Mail Item / Outlook Express / IMAP - Read Item without
Graph 12 : Generic Mail Item / Outlook Express / IMAP - Read item with
Graph 13 : Generic Mail Item / Outlook Express 5.01 / POP - Send item without
Graph 14 : Generic Mail Item / Outlook Express / POP - Send item with
Graph 15 : Generic Mail Item / Outlook Express 5.01 / POP - Read Item without
Graph 16 : Generic Mail Item / Outlook Express 5.01 / POP - Read Item with
Graph 17 : Generic Mail Item / Netscape Messenger IMAP - Send item without
Graph 18 : Generic Mail Item / Netscape Messenger IMAP - Send item with
186
Graph 19 : Generic Mail Item / Netscape Messenger IMAP - Read item without
attachment (comparative: HTML, Plain Text)................................................ 70
Graph 20 : Generic Mail Item / Netscape Messenger IMAP - Read item with
Graph 21 : Generic Mail Item / Netscape Messenger POP - Send item without
Graph 22 : Generic Mail Item / Netscape Messenger POP - Send item with
attachment(comparative: HTML, Plain Text)................................................. 74
Graph 25 : Generic Mail Item - Send item in RTF format without attachment
(comparative: Outlook 2000, Outlook 97)...................................................... 80
Graph 26 : Generic Mail Item - Send item in RTF format with attachment
Graph 27 : Generic Mail Item - Send item in HTML format without attachment
(comparative: Outlook 2000, Outlook Express IMAP / POP, Netscape IMAP /
POP, OWA) .................................................................................................... 81
Graph 28 : Generic Mail Item - Send item in HTML format with attachment
POP, OWA) .................................................................................................... 82
Graph 29 : Generic Mail Item - Send item in Plain Text format without attachment
(comparative : Outlook 2000, OExpress IMAP / POP, Netscape IMAP / POP)
........................................................................................................................ 82
Graph 30 : Generic Mail Item - Send item in Plain Text format with attachment
POP) ............................................................................................................... 83
Graph 31 : Generic Mail Item - Read item in RTF format without attachment
Graph 32 : Generic Mail Item - Read and Open Item in RTF format with
attachment (comparative: Outlook 2000, Outlook 97) ................................... 84
Graph 33 : Generic Mail Item - Read item in HTML format without attachment
POP, OWA) .................................................................................................... 84
Graph 34 : Generic Mail Item - Read and Open item in HTML format with
attachment (comparative: Outlook 2000, Outlook Express IMAP / POP,
Netscape IMAP / POP, OWA) ....................................................................... 85
Graph 35 : Generic Mail Item - Read item in Plain Text format without attachment
POP) ............................................................................................................... 85
Graph 36 : Generic Mail Item - Read and Open item in Plain Text format with
attachment (comparative: Outlook 2000, Outlook Express IMAP / POP,
Netscape IMAP / POP)................................................................................... 86
Graph 37 : Calendar actions (comparative : Outlook 2000, Outlook 97, OWA).... 92
Graph 38 : Contact actions (comparative : Outlook 2000, Outlook 97, OWA)...... 92
Graph 39 : Task actions (comparative : Outlook 2000, Outlook 97)...................... 93
Graph 40 : Public Folder / Outlook 2000 - Send Item without attachment

(comparative: RTF, HTML, Plain Text) .......................................................100
Graph 41 : Public Folder / Outlook 2000 - Send Item with attachment
Graph 42 : Public Folder / Outlook 2000 - Read item without attachment
Graph 43 : Public Folder / Outlook 2000 - Read item with attachment
Graph 44 : Public Folder / Outlook Express / NNTP - Send Item without
attachment(comparative: HTML, Plain Text)...............................................107
Graph 45 : Public Folder / Outlook Express / NNTP - Send Item with attachment
(comparative: HTML, Plain Text) ................................................................108
Graph 46 : Public Folder / Outlook Express / NNTP - Read Item without
attachment(comparative: HTML, Plain Text)...............................................108
Graph 47 : Public Folder / Outlook Express / NNTP - Read Item with attachment
(comparative: HTML, Plain Text) ................................................................109
Graph 48 : Public Folder / Netscape Messenger / NNTP - Send item without
attachment (comparative: HTML, Plain Text)..............................................112
Graph 49 : Public Folder / Netscape Messenger / NNTP - Send item with
Graph 50 : Public Folder / Netscape Messenger / NNTP - Read item without
Graph 51 : Public Folder / Netscape Messenger / NNTP - Send Item with
Graph 52 : Public Folder- Send item in HTML format without attachment
(comparative : Outlook Express, Netscape)..................................................115
Graph 53 : Public Folder- Send item in HTML format with attachment
Graph 54 : Public Folder- Send item in Plain Text format without attachment
Graph 55 : Public Folder- Send item in Plain Text format with attachment
(comparative: Outlook Express, Netscape)...................................................117
Graph 56 : Public Folder- Read item in HTML format without attachment
Graph 57 : Public Folder- Read item in HTML format with attachment
Graph 58 : Public Folder- Read item in Plain Text format without attachment
Graph 59 : Public Folder- Read item in Plain Text format with attachment
188
17.2. Figures
Figure 1 : Network Topology ................................................................................... 9

Figure 2 : Exchange 2000 installed components (1/2)............................................ 12
Figure 3 : Exchange 2000 installed components (2/2)............................................ 12
Figure 4 : Terminal service configuration .............................................................. 13
Figure 5 : Terminal Services Client........................................................................ 17
Figure 6 : Network Monitor Capture Filter ............................................................ 18
Figure 7 : Network Topology ............................................................................... 139
Figure 8 : Exchange 2000 installed components .................................................. 142
Figure 9 : Routing Groups Configuration............................................................. 142
Figure 10 : Front-End Switch ............................................................................... 143
Figure 11 : Recipient Update Services View........................................................ 145
Figure 12 : FE/BE Message Sending Traffic Frames Overview .......................... 150
Figure 13 : FE/BE Message Reading Traffic Frames Overview .......................... 157
Figure 14 : Paris Routing Group Configuration (1/2)........................................... 159
Figure 15 : Paris Routing Group Configuration (2/2)........................................... 160
Figure 16 : Nantes Routing Group Configuration (1/2)........................................ 160
Figure 17 : Nantes Routing Group Configuration (2/2)........................................ 161
Figure 18 :Routing Group Message Sending Traffic Frames Overview .............. 168
Figure 19 : DSAccess Overview........................................................................... 172
Figure 20 : First discovery of the GC by Outlook 2000 client ............................. 182
Figure 21 : DSProxy Access................................................................................. 184
The information contained in this document represents the current view of Microsoft Corporation on the issues
discussed as of the date of publication. Because Microsoft must respond to changing market conditions, it should
not be interpreted to be a commitment on the part of Microsoft, and Microsoft cannot guarantee the accuracy of
any information presented after the date of publication.
This White Paper is for informational purposes only. MICROSOFT MAKES NO WARRANTIES, EXPRESS OR
IMPLIED, AS TO THE INFORMATION IN THIS DOCUMENT.
Complying with all applicable copyright laws is the responsibility of the user. Without limiting the rights under
copyright, no part of this document may be reproduced, stored in or introduced into a retrieval system, or
transmitted in any form or by any means (electronic, mechanical, photocopying, recording, or otherwise), or for
any purpose, without the express written permission of Microsoft Corporation.
Microsoft may have patents, patent applications, trademarks, copyrights, or other intellectual property rights
covering subject matter in this document. Except as expressly provided in any written license agreement from
Microsoft, the furnishing of this document does not give you any license to these patents, trademarks, copyrights,
or other intellectual property.
Unless otherwise noted, the example companies, organizations, products, domain names, e-mail addresses,
logos, people, places and events depicted herein are fictitious, and no association with any real company,
organization, product, domain name, email address, logo, person, place or event is intended or should be
inferred.
© 2001 Microsoft Corporation. All rights reserved.
Microsoft, Active Directory, Windows, and Windows NT are either registered trademarks or trademarks of
Microsoft in the United States and/or other countries.
The names of actual companies and products mentioned herein may be the trademarks of their respective
owners.

Client Network Traffic With Exchange 2000

Enviado por

Dados do documento

Descrição original:

Direitos autorais

Formatos disponíveis

Compartilhar este documento

Compartilhar ou incorporar documento

Opções de compartilhamento

Você considera este documento útil?

Este conteúdo é inapropriado?

Direitos autorais:

Formatos disponíveis

Client Network Traffic With Exchange 2000

Enviado por

Direitos autorais:

Formatos disponíveis

Client Network Traffic with Exchange 2000

Client Network Traffic

Microsoft Consulting Services – France

The example companies, organizations, products, people, and events depicted

4.3.4. Outlook Express Measurements Analysis 28

6.10. Mail Item Analysis 80

13.1.5. Exchange 2000 configuration 141

Microsoft Consulting Services, France

FR-PSS REC Biz Apps

Deployment Assets Team

2. Document Vision / Scope

Deploying Microsoft Exchange 2000 Server in a company requires a thorough

3. Test Lab Configuration

3.1. Lab design

Figure 1 : Network Topology

WINDOWS 2000 Server

WINDOWS 2000 Server

WINDOWS 2000 Professional

WINDOWS 2000 Server Lonbdon 06

3.2. LAN design

Table 1 : Server configurations

Server IP address IP Subnet mask Role

3.3. Servers Characteristics

Table 2 : Server characteristics

Server CPU Memory OS Other

3.4. Client Characteristics

Table 3: Client characteristics

Client CPU Memory Operating Other

3.5. Exchange 2000 Configuration

Figure 2 : Exchange 2000 installed components (1/2)

Figure 3 : Exchange 2000 installed components (2/2)

3.6. Terminal Services Configuration

The LONDON-08 server is running Terminal Services and Terminal Services

3.7. Client Configuration

3.7.1. Outlook 2000 Client Configuration

3.7.2. Outlook 97 Client Configuration

The folder list is enabled for the tests on public folders.

3.7.3. Outlook Express Client Configuration

3.7.3.1. POP3 Account Settings

3.7.3.2. IMAP4 Account Settings

3.7.3.3. Lightweight Directory Access Protocol (LDAP) Account

3.7.3.4. Network News Transfer Protocol (NNTP) Account

3.7.4. Netscape Messenger Client Configuration

The following options are enabled:

3.7.4.1. POP3 account settings

3.7.4.2. IMAP4 account settings

3.7.4.3. NNTP account settings

3.7.5. Outlook Web Access Client Configuration

3.7.6. Terminal Services Client Configuration

Figure 5 : Terminal Services Client

3.8. Measurement Methodology

A static network address resolution was established prevent unexpected network

ARP.EXE –s 10.0.0.1 01-02-03-qb-cd-ef-99

• Messages with attachments: had Microsoft Word or Microsoft PowerPoint files

• Public folders tests and NNTP access:

• Messages with attachments: Microsoft Word or Microsoft PowerPoint files of

4. Log On and Log Off

4.1. Tests Performed

4.2. MAPI Clients: Microsoft Outlook 2000,

For MAPI clients, the generated traffic depends on various parameters:

4.2.1. Test Details

To capture the traffic, the user name was not resolved.

LOGON2: First startup, existing mailbox, new profile