Você está na página 1de 7

Literature Synthesis

A Document presented to the

Department of Computer Science


at the

University of Cape Town

by
Pheeha Machaka(MCHPHE002)
WiFi spots Monitoring Tools: Databases
Pheeha Machaka, MCHPHE002
Honours Laboratory, Computer Science Department
University Of Cape Town
Rondebosch, 7701, South Africa
pheeham@gmail.com
+27 83 593 3793

ABSTRACT even at public places of congregation such as conference


venues, airports, shopping malls, hotels, libraries, arenas
Wifi has become widely used in public areas like airports, and so on – places where they spend a considerable amount
coffee shops, and university campuses; and companies like of time outside private networks. (Anand Balachandran,
RedButton has taken advantage of this high demand for
2005)
wireless access to internet. They now want a network-
monitoring tool that will help them monitor and make
strategic business decision based on the networks
performance. In this paper I will investigate the best
database tool that will be used for the monitoring tool that
we will be develop. This paper is focused on the data
storage module of the project. We will explore related work
in this field, and we will look at existing databases and
compare their performances. We will also look at available
tools that are used for monitoring networks.

Categories and Subject Descriptors


C.2.3 Network Operations: Network Management, Network
Monitoring, and Public Network

General Terms
Management, Performance, Design, Security Figure 1. An illustration of
how WiFi works
As illustrated by Figure 1(picture taken from
Keywords www.howstuffworks.com), the AP can be a “router-like”
Wireless Fidelity (WiFi); Access Points (AP); Protocol, device, that will broadcast a signal within certain radius. A
Framework, Monitoring Tool, Databases laptop or a cell phone PDA that is capable of receiving the
802.11 standard network signal will be able to use the AP
for access to the world wide web.
1. INTRODUCTION The latest laptops are equipped with devices that can detect
Wireless Fidelity (WiFi) is a wireless networking a wireless signal(s) that’s being broadcast in that particular
technology that uses radio waves to provide high-speed, area by the available AP(s). If the laptop computer does not
wireless internet and network connections. This technology have a built-in wireless LAN detecting device, the user can
is based on the IEEE 802.11 standards, including the insert a wireless adapter that can be plugged into the PC
802.11a and 802.11b (IEEE, 1999). This technology has card slot or the USB port.
proven to be the fast-wireless networking approach that is Over the past few years, there has been an increase in the
relatively easy and inexpensive to implement (Steven J. number of wireless users. These users have access to the
Vaughan-Nichols, 2003). This is made possible by using internet at various places. Companies like RedButton have
wireless access point (AP) or hotspots. The AP will taken advantage of the increasing demand for the use of
broadcast the signal to WiFi-capable devices within the wireless access to the internet.
AP’s range. Capable devices can be, from laptops to
RedButon have placed their WiFi spots devices in coffee
cellular mobile devices and PDA’s. shops, hotels, corporate offices, and many other places that
AP’s are installed in business settings. However, business have a potential demand for wireless internet access. The
networking no longer happens just in offices (Steven J. increase in the demand for wireless access to internet in
Vaughan-Nichols, 2003). Today, widely travelling laptop various places, has led to an increase in the number of WiFi
users access the internet at a variety of places and hotpots that RedButton has installed, in these various
environments including their homes, corporate offices, and strategic business positions. As the number of hotspots
increase, it becomes more difficult to monitor the devices.  Bandwidth usage broken down by MAC address
In this project, we will design a WiFi Hotspot Network  Broadcasts as a percentage of total packets
Monitoring Tool.
 Packet loss and error rate
Therefore, in this paper, we will thoroughly investigate the
various monitoring tools that are available, and will
particularly look at the open source monitoring tools.
There are three parts to this project. The first one being the Internet Statistics
protocol (or framework) to be used for collecting the data  Proxy server cache hits
from hotspot devices; the second being the database tool to
be used for storing, retrieving and managing the data  Most frequently visited sites
collected; the third is the tool to be used for presenting  Internet bandwidth usage by users
(visualization) the data collected to the systems
 Status of backups
administrator.
My focus in this project will be on the various database
tools available to store monitoring data collected from the System Health statistics
hotspots that RedButton has installed in the country; while  Memory usage
my project partners, Khathutshelo Mufamadi and Kuo-An
 Swap file usage
Wang, will focus on the frameworks and visualization,
respectively.  System load
 Temperature, fan speed and system voltage

2. Network Monitoring (Chapter 6: Security and Monitoring: Open Source


Book on Wireless Internet for Developing Worlds)
When monitoring a network, one will use logging and
analysis tools that accurately determine traffic flows,
utilization, and other performance indicators of the
network. The monitoring tools will be used to give 2.2 Middleware
numerical and graphical representations of the state of the
network. The monitoring tool is meant to help the system
administrator to answer questions like: The gap between application layer and lower layer
What are the most popular services used in the network? protocols can cause performance degradation and
Which WiFi Hotspot is used heavily? Which other wireless implementation inefficiencies in mobile wireless networks.
channels are in use in a particular area? What websites do This problem can be minimized if a layer supporting
users visit most frequently? At what time of the day; the coordination between the upper and lower layer protocols is
month; and even the year is the network most utilized? Is created to mediate between these layers. Middleware
the amount of inbound and outbound traffic close to the enables communication between the network operating
network’s capacity? Are there indications of unusual system and applications by providing application
network activities that are consuming bandwidth? programmers with high-level abstraction (Meiso K. Denko,
2009).
In this project, we will build a middleware for the network-
2.1 What are we monitoring? monitoring tool. The network we are monitoring is made up
of different AP devices from different vendors; which use
When monitoring a network, one should have a set of
different protocols and platforms to manage the devices.
statistics that they would want to collect from the network’s
We want the devices and applications to communicate
activities. The systems administrator should have key
performance indicators (KPI) that they would want to across different platforms.
measure and control. (Chapter 6: Security and Middleware will ease the transformation of markup
Monitoring: Open Source Book on Wireless Internet for languages, the delivery of content and data, protocol and
Developing Worlds) There are many KPI’s that can be device recognition, and the adaptation of data formats for
used for measuring the network’s performance; here is a compatibility with multiple databases (Vaughan-Nicholis,
small list: 2004).

Wireless Statistics
 Received signal and noise from all AP’s
 Number of associated stations
 Detected adjacent networks and channels
 Excessive retransmission

Switch Statistics
 Bandwidth usage per switch port
relational database (SolidTech – provided by SOLID
relational databases, a product by IBM –
www.solidtech.com)

(Oiteker, 1998) of the Swiss Federal Institute of


Technology wrote a paper describing the history of the
Multi Router Traffic Grapher (MRTG). The paper
described the old and current versions of MRTG. The first
version of the MRTG program was a Perl script which uses
external utilities to do SNMP queries and to create images
for display in HTML. It logged its data to an ASCII file.
The logfiles were then used to draw graphs on the web
page. The logfiles were in plain ASCII, were each line
started with a time stamp followed by the corresponding
traffic data. The third version of MRTG used the Round
Robin Database (RRD). With this storage mechanism,
MRTG became faster and more configurable. MRTG has
moved from simple, plain files’ storage mechanism to a
Figure 2. An illustration of more sophisticated and reliable database storage
how middleware works mechanism.
This shows that simple ASCII files can be used for network
Middleware allows AP devices, application server, and monitoring, but they have greater performance
database systems to work together as a single application. disadvantages. (Oiteker, 1998)
The middleware identifies the devices and the
communications protocol it uses to the rest of the system,
letting all parties know how to talk to each device. The 4. Current Situation
middleware gives the database and device a transparent
route to each other’s data, and it provides the application
server with an API for working with the device (Vaughan-
Nicholis, 2004).

3. Related Work
(Marika K. Thottan, 2003) from Lucent technologies Inc.
developed a Monitoring tool (SEQUIN) for a Multiprotocol
Label switching (MPLS)-based network. An MPLS-based
network allows packet switching at a high rate, while
retaining the flexibility of Internet Protocol (IP). It enables
Service Providers to preserve network feature where IP can
provides support for Quality of Service (QoS) through
differentiated services. More details on how this type of
network functions can be found in (Marika K. Thottan, Figure 3. A snapshot of a view of current
2003). To ensure that the level of QoS is met, service system from the meeting with RedButton
providers need to monitor the network and keep track of
RedButton does not have a framework for collecting the
Bandwidth utilization and other QoS metrics. SEQUIN
data, or probing the wireless devices. They would want to
uses SNMP-based techniques to keep track of QoS of
have a framework were data can be collected from all the
metrics for a network service provider. This monitoring
AP devices in the network. The framework’s daemon tool
tool had the same modules as the ones we are planning to
will have access to the AP devices through a OpenVPN
have in our project. The modules different tasks included
tunnel.
network polling, computation of QoS Metrics, and
visualizations. They would like to have the data stored to a central server
for historic reporting. This data must therefore be computed
In their monitoring tools database structure, they classified and represented in a format that will allow decision-
data stored in the database as static or dynamic. Static
making.
information includes system configuration, network
element configuration, network topology, and monitoring
agent information. Dynamic information consists of polled
SNMP data and information computed with this data. The
polled data were kept in tables that were cleared frequently,
while tables with QoS statistics where kept for a longer
period as they were used for decision making. They used a
5. Monitoring Tool Cacti (http://www.cacti.net)
Redbutton would like to have a tool that will monitor how Cacti is a front-end for RRD Tool, it harnesses the power of
the network is utilized over a long period of time; therefore RRD Tool data storage and graphing functionality. The
they would like to have a Trending Tool. Trending Tools front-end is written in PHP. It does the work of
work by periodically monitoring the network’s activities; maintaining graphs, data sources, and handles the actual
collecting and analyzing the data, then representing the data data gathering, as illustrated in Figure 3(picture from
in a numerical and graphical format. linux.softpedia.com). To handle data gathering, you can
feed cacti the paths to any external script/command along
with any data that the user will need to "fill in", cacti will
5.1 Device probing then gather this data and populate a MySQL database/the
round robin archives. There is support for SNMP devices,
In this project, we will have to probe or retrieve data from
and custom scripts can easily be written to poll any
the AP devices. My project partner, Khathutshelo
conceivable network event.
Mufamadi, will suggest the tool and protocol that will be
used for this part of the project.

5.2 Visualization
The result will need to be presented in a human readable
form. This part of the project will be conducted by my
other project partner, Kuo-An Wang. He will use
visualization tools to represent the data collected from the
AP devices. These tools will represent the data using
numerical data and graphs.

5.3 The database tool


In this project, I will focus on the database tools that will be Figure 4. A snapshot of Cacti Monitoring Tool
used to store the data. This will offer a platform where data
from different router’s and AP devices will be stored and OpenNMS(http://www.opennms.org)
managed.
OpenNMS is the world's first enterprise grade network
RRDTool(http://oss.oetiker.ch/rrdtool/) management platform developed under the open source
Since we will be modeling data from a network over a long model. It works across all platforms. It is written in Java. It
period of time, the Round Robin Database (RRD) Tool can also collect, store and report on data collected from the
proves to be the most widely used industry standard, high AP devices using SNMP. The tool can evaluate the
performance data logging and graphing system for time network’s data or performance against thresholds.
series data.
The RRD Tool assumes time-variable data in intervals of a
certain length (time steps). Because data may not always be
available at just the right time, the RRD Tool will
interpolate any submitted data to fit its internal time-steps
(Urheber, 2005) (Oiteker, 1998).

PostgreSQL(http://www.postgresql.org)
PostgreSQL is a powerful, open source object-relational
database system. It has a strong reputation for reliability,
data integrity and correctness. It runs across all major
operating systems. It can run stored procedures in major
programming languages.
Figure 5. A snapshot of the OpenNMS Monitoring
Tool.
5.4 The Monitoring Tools The above mentioned tools are very well designed for
There are several network monitoring tools that use the monitoring network activities. Therefore, I chose the tool
RRD Tool as their DBMS. I chose the most widely used that will be best suited for this project, and below is a table
networking monitoring tools that use RRD Tool, and (Table 1) comparing the tools; and which one will be best
compared them. to use. I compared the tools using the following parameters
or ability to perform the functions: service polling, data
collection, data storage, trap receiver and thresholding.
Network Model
Table 1: Compares the Monitoring Tools chosen The network model was developed after realizing the
Hierarchal Model’s shortcomings. There was a need for a
OpenNMS Cacti more natural modeling of data, instead of the one-to-many
Service Polling yes yes modeling. This model allowed a many-to-many
relationship of data. Data in this model is represented as
Data Collection SNMP, syslog SNMP, scripts records and sets. A set consists of an owner record type, a
Data Storage PostgreSQL, RRD RRDTool set name, and a member record type. An owner record type
Trap Receiver SNMP, syslog SNMP, scripts can also be a member or owner in another set. A member
record type can have that role in more than one set, this
Thresholding yes yes allows for a many-to-many relationship. In this model
From the above table, it is evident that the best tool to use redundancy of data is represented more efficiently (Codd,
is Cacti, which is the front-end for RRDTool, and creates 1980),
graphs in a MySQL database.
Relational Model
6. Databases This model was introduced by E.F. Codd in 1970. It allows
a definition of data structures, storage and retrieval
6.1 Database Models operations and integrity constraints. There are three key
There exist several types of databases models, and we will terms in this model: relations, attributes and domains. A
look at how they differ from the others. relation is a table with columns and rows (called tuples).
Flat Model The named columns of the relation are called attributes, and
the domain is the set of values the attributes are allowed to
Flat file models databases are simply plain-text files or take. In a table, a single attribute or a set of attributes can
binary files; where data is represented in a two-dimensional act as a key (primary key) to uniquely identify each tuple in
array format (using columns and rows). the table (Codd, 1980).
Advantages
 Very simple to create and access Object database models
 Contain lots of redundant information
The object database model is similar to that of the
Disadvantages Relational Model, but using the object-oriented paradigm.
 The redundant information makes the search In this database model key ideas of object-oriented
process very slow programming are produced, like: classes, inheritance,
encapsulation and polymorphism. This model tries to bring
 Very inefficient together the world of databases and that of object-oriented
(Fowler) programming (Atkinson, 1995).

Hierarchal Model 6.2 Open Source Database Management


Data in this model is stored in a tree-like structure. The Systems
relationship between entities in this model can be best
described using the parent-child relationship; where a There are many open source database management systems
parent can have many children and a child has only one available, but we will now look at only a few of them and
parent. compare their features.
Advantages In April 26, 2005, Herausgeber und Urheber and his
colleagues from FabaLabs Software, together with the
 Easy to implement Institute for Applied Knowledge Processing (AFW),
 Data is organized very well published a research paper comparing the enterprise
 Easy to master functionalities of open source database management
Disadvantages systems. They evaluated the following DBMS’: FireBird
1.5.2, Ingres r3 3.0.1, MaxDB 7.5.0.23, MySQL 4.1.10 and
 It forces representation of data into a one-to- PostgreSQL 8.0.1. These may be older versions of the
many relationship only DBMS, but the data and results may help in deciding which
 Manipulation of data using other operators DBMS to implement for the system. From the results they
becomes complex obtained, we will look at the limits and ACID compliance
 Makes delete entries very difficult (Urheber, 2005).
 Searching in this model is very difficult
(Lochovsky, 1976) Limits per Database
When choosing a database, one will need to carefully look
at the limitations that the database may have. There are
limits for databases like maximum database size, maximum Chapter 6: Security and Monitoring: Open Source Book on
number of file, maximum file/volume size, maximum Wireless Internet for Developing Worlds.
number of table and many other limitations (Urheber, Codd, E. (1980). Data Models in Database Management.
2005).
David L. Wells, J. A. (1992). Architecture of an Open
Table 2: Comparison of the DBMS’ limits Object-Oriented Database Management System.
Fowler, G. (n.d.). A Flat File database Query Language -
cql.
IEEE. (1999). Wireless LAN Medium Access Control
(MAC) and Physical Layer (PHY) specifications.
Lochovsky, D. C. (1976). Heirarchical database
management: a survey.
Marika K. Thottan, G. K. (2003). SEQUIN: An SNMP-
ACID compliance Based MPLS Network Monitoring System.
We will now look at how each of the DMBS’ handles Meiso K. Denko, E. S. (2009). Enhanced cross-layer based
transactions. We will want to find out if they are ACID middleware for mobile and ad hoc networks.
compliant (Urheber, 2005). Oiteker, T. (1998). The Multi Router Traffic Grapher.
Atomicity: A transaction is an atomic entity, which will Steven J. Vaughan-Nichols, I. X. (2003). The Challenge Of
either entirely succeed or fail. Wifi Roaming.
Consistency: A transaction should transform the database Stonebraker, M. (n.d.). Object-Relational DBMS - The next
from one consistent state into another. Wave.
Isolation: A transaction must not influence other Urheber, H. u. (2005). Comparison of the Enterprise
transactions in any way. functionalities of Open Source Database Management
Durability: Committed transactions should remain durable, Systems.
even in case of a system failure. Vaughan-Nicholis, S. J. (2004). Wireless Middleware: Glue
for the Mobile infrastructure.
All the DBMS’ that were evaluated in their research paper
were ACID compliant.

7. Conclusion
In this project, we will use the relational database system
that is used by Cacti, RRDTool. The firmware’s data can be
collected using a syslog method or the Simple Network
Management Protocol (SNMP) method. The choice
between the methods will depend on the bandwidth needed
and the storage space needed to perform the methods, for
data collection from the AP devices.
We will design data tables that are needed for the system,
determine the fields needed for the tables and the
relationships between the tables. This tool will then
populate the database tables with the data captured from the
firmware.
The visualization component of the network-monitoring
tool, will interpret the data collected and stored in the
database; the component will produce numeric and
graphical data for the systems administrator to react and
make decision based on the data presented.

8. References
Anand Balachandran, G. M. (2005). Wireless Hotspots:
Current Challenges and Future Directions.
Atkinson, M. (1995). The object-Orinted Database System
Manifesto.

Você também pode gostar