Você está na página 1de 12

Expert Reference Series of White Papers

Fundamentals of SCSI and iSCSI

1-800-COURSESwww.globalknowledge.com

Fundamentals of SCSI and iSCSI


Raymond B. Dooley, Global Knowledge Instructor, CCSI, CCNP, CCDP

Introduction
The computer world is full of buses. Most people are familiar with the universal serial bus (USB), which works ne for storing photos, and music les, but it is not big enough to support the types of le and data retrieval functions needed in a corporate data center. To support the retrieval of stored data from a disk, group of disks, optic memory array, or tape drives, another type of bus and protocol is required. Small Computer Systems Interface (SCSI) was originally a proprietary design by Shugart Associates in the 1970s. In 1986 the SCSI (pronounced scuzzy) standard was released by the American National Standards Institute (ANSI), and since then it has evolved through SCSI-1 and SCSI-2, to the current SCSI-3. This white paper will focus on: Protocol Overview and Role Parallel Bus Technology Architecture Model SCSI over Fibre Channel SCSI Operations Commands and Messages Building an I/O Request Internet SCSI (iSCSI)

Copyright 2011 Global Knowledge Training LLC. All rights reserved.

SCSI Protocol Overview and Role

Figure 1.

SCSI denes the relationship between initiators (hosts) and targets (disk or tape). The SCSI-3 application resides in the host and provides middle-layer OSI Model (session and transport) services for upper layer file management protocols. Several components and options are dened: The Software initiator uses code in the OS (kernel-resident device driver) which uses the existing infrastructure. This is the most common type. The Hardware initiator uses dedicated hardware, typically in combination with rmware running on the hardware to implement SCSI. The hardware is the Host Bus Adapter (HBA), and is more common for servers in data centers. SCSI targets include the following: A single disk Just a Bunch of Disks (JBOD) RAID - Redundant Array of Independent Disks Tape Storage Optic Memory

Copyright 2011 Global Knowledge Training LLC. All rights reserved.

Each individual disk is identied in SCSI as a Logical Unit Number (LUN). SCSI is fast (320 Megabytes per second), but limited to 25 meters of cable.

Parallel Bus Technology


The SCSI parallel bus is either narrow (8 bits) or wide (16 bits), with wide used more often in modern systems. Data is sent simultaneously over multiple wires (in parallel, see Figure 2 below).

Figure 2

SCSI is half-duplex meaning data travels in one direction at a time, and only one device can control the bus at one time. A device must assume exclusive control over the bus in order to communicate. Another name for this is simplex communication (Like a Western Union telegram). The SCSI initiator selects the SCSI target and sends a command to initiate data transfer. At the end of the transfer, the bus is deselected and the bus is free. Data bits and control bits are sent on separate wires; this is called out of band control. Each SCSI device has a unique Identifier (ID). The initiator is always ID = 7, which is the highest priority. In a SCSI target, in addition to the ID, a LUN then identifies the specific target disk. The process requires an exchange of commands to and from the two addresses, then the transfer of data.

SCSI-3 Architecture Model


Figure 3 represents the SCSI-3 architecture model. The model consists of four layers of functionality (from the bottom of the graphic working up, with number one at the bottom of the graphic and number four at the top).

Copyright 2011 Global Knowledge Training LLC. All rights reserved.

Figure 3.

The physical interconnect layer denes the characteristics of the link between initiator and target. FC is fibre channel Serial Storage Architecture (SSA) is common in data centers where scalability is an issue IEEE 1394 is the Firewire specification SCSI Parallel Interface (SPI) is in every PC (on the far left side of the graphic) The transport protocol layer denes session management, which is techno-speak for the rules of the conversation. SCSI-FCP for fibre channel Serial Storage Protocol (SSP) for SAS devices Serial Bus Protocol (SBP) for IEEE 1394 The shared command set layer consists of command sets for accessing storage resources (disk arrays, optic arrays, etc.). The SCSI Common Access Method (CAM) defines the SCSI drive application programming interface (API). APIs are used by application program developers to insure that the application written has proper access to storage.

Copyright 2011 Global Knowledge Training LLC. All rights reserved.

In a medium-to-large enterprise data center, there will be a bre channel fabric. The SCSI initiator and target will communicate over this fabric. SCSI commands, data, and responses are carried in the payload of bre channel frames. Fibre channel can be described as a very fast pipe, which never loses any data or becomes congested, that provides a conduit for the SCSI commands and trafc. In SCSI-FCP, the SCSI IDs are mapped to the unique worldwide name in each Fibre Channel port. This is what makes the pipe work.

SCSI Operations
SCSI species three phases of operation: 1. Phase One is the Command Phase. In this phase the required commands and parameters are sent. 2.  Phase Two is the Data Phase. The transfer of data in accordance with the command occurs in this phase. 3. Phase Three is the Response Phase. Confirmation of command execution is received in this phase.

Figure 4.

Each SCSI command is formatted as a Command Descriptor Block (CDB). The majority of the SCSI protocol is controlled by the initiator, while the target is passive and waits for a command. Only the initiator can initiate a SCSI operation by selecting a SCSI target and sending a CDB to it. If the CDB contains a Read command, the SCSI target moves its heads into position and retrieves the data from disk sectors. If the CDB contains a Write command, the SCSI target prepares its buffers and returns an indication of ready. When the initiator receives the ready, the data is written to disk.

Copyright 2011 Global Knowledge Training LLC. All rights reserved.

When the operation is complete, the SCSI target returns a response to indicate a successful, or unsuccessful, data transfer.

SCSI Commands
SCSI supports several specic commands for each media type, and primary commands that all devices understand. The Session Layer (OSI layer 5) protocols are complex and require more command-level functionality than any other layer, with the possible exception of layer 7 (Application). Since SCSI performs session layer functions for data storage and retrieval, it is also complex. Table 1 shows a few of the many available commands, formatted as CDBs: Command Inquiry Report LUNs Test unit ready Report capacity Write buffer Read buffer
Table 1.

Function What device are you? How many LUNs do you have? Is the LUN available? What size is each LUN? Write data to target LUN. Read data from target LUN.

SCSI Messages are an additional way for the initiator and target to communicate. Some SCSI transmission parameters are not tied to a specic command but to the initiator-target relationship. Several parameters, including transfer speed, data width, abort, and restore pointers, are established between initiator and target to maintain accurate information exchange.

Building an I/O Request


Figure 5, below, provides an example of an initiator talking to a target.

Copyright 2011 Global Knowledge Training LLC. All rights reserved.

Figure 5.

The process breaks down into the following steps: 1. The application makes a file I/O request to Volume Manager 2. Volume Manager maps volume to SCSI ID and target LUN 3. File System maps files to blocks, makes block I/O request 4. Command, LBA (Logical Block Address), block count, and LUN sent to SCSI driver 5. SCSI driver creates CDB 6. FC (Fibre Channel) driver creates command frame with CDB in payload 7. FC driver sends command frame to target LUN and awaits response As errors may occur in this process, the SCSI Retry command causes the entire operation to be repeated.

Internet SCSI (iSCSI)


SCSI specications limit data transfer distance to 25 meters, but many data center design parameters require more distance between devices. iSCSI (Internet SCSI) allows initiator and target to exchange SCSI commands

Copyright 2011 Global Knowledge Training LLC. All rights reserved.

using IP networks over TCP ports 860 and 3260. Since iSCSI requires no dedicated cabling (fibre channel) it is often viewed as a low-cost alternative to bre channel (see Table 2 on page 11 for other comparison points).

Figure 6.

There are two major applications for iSCSI: 1.  Storage Consolidation - Organizations move disparate storage resources from servers around their network to central locations (data centers). This allows for more efciency in the allocation of storage. 2.  Disaster Recovery - Organizations mirror storage resources from one data center to a remote data center, which can serve as a hot standby in the event of a prolonged outage. iSCSI SANs allow entire disk arrays to be migrated across a WAN with minimal conguration changes, in effect, making storage routable in the same manner as network trafc.

Figure 7

An interface from the SCSI device to the iSCSI world (network) is required. The simplest of these is the standard NIC (Network Interface Card). However, as volumes increase, the TCP/IP processing across the interface will overload the host. Another option is to build TCP/IP processing on the NIC, called a TCP Offload Engine (TOE). A TOE can be installed as partial offload (everything but error recovery) or full offload. To achieve maximum performance, the iSCSI HBA (Host Bus Adapter) is used.

Copyright 2011 Global Knowledge Training LLC. All rights reserved.

Figure 8.

SCSI standards dene a client server relationship between the initiator and target. iSCSI denes these as the Network Entity. The network entity contains an iSCSI node (either initiator or target), which is identied by an iSCSI node name. There are three formats: 1.  iSCSI Qualied Name (IQN) - It is a month and year date followed by a reverse domain name. (2010-04.com.acme:storage.tape.sys1.xyz) 2. IEEE Extended Unique Identier (EUI) - 64 bits 3. T11 Network Address Authority (NAA) - 64 or 128 bits IQN names are the most commonly used. If the target node is a storage array, it can contain one or more SCSI LUNs. iSCSI initiator nodes communicate with iSCSI target nodes through network portals, identified by IP address. Network portals can be wireless ports. Table 2 provides a quick comparison of bre channel and iSCSI as a SAN solution. What it does not show is that fibre channel can also run over IP (FCIP), which makes the choices even more challenging. One of the major disadvantages of iSCSI in the data center is meeting the scalability and performance required in high-volume environments.

Copyright 2011 Global Knowledge Training LLC. All rights reserved.

10

Fibre Channel Designed for enterprise markets High bandwidth of 1 Gbps, 2 Gbps, 4 Gbps Low latency of 2 msec per port Large payload of up to 2112 bytes Low overheads of 5.5% for 1KB payload, 3% for 2KB payload Short distance of 10 km per link for single mode ber High cost HBAs and switch ports
Table 2

iSCSI Designed for SMB markets Low bandwidth of 1 GigabitEthernet, 10 GigabitEthernet High latency; IP has msec latency Smaller payload of up to 1500 bytes; up to 9000 bytes using jumbo frames Higher overheads of 8% for 1KB payload Long distance; no theoretical limit over IP networks, but high latency Low cost: use existing NiC and LAN; iSCSI HBAs are expensive

Another issue with all SAN solutions is network security. iSCSI supports authentication with CHAP (Challenge Handshake Authentication Protocol) but it must be properly implemented to avoid known vulnerabilities. Also, iSCSI is sent in clear text, and while IPSec may be used for encryption, it has presented problems when used with iSCSI. Server Message Encryption (SME) can be used in non-SCSI environments.

Summary
To recap, SCSI is present in all personal computers and has been a standard for le storage and retrieval for many years. It defines an initiator/target client/server relationship which includes addresses, commands, and specied operations. The SCSI model includes numerous interfaces including TCP/IP, fibre channel, directly attached parallel bus, and firewire. The many SCSI commands are formatted as CDBs. Fibre Channel and iSCSI are the two major choices for building out SCSI over a network. Both are present in modern data centers.

Learn More
Learn more about how you can improve productivity, enhance efciency, and sharpen your competitive edge through training. Data Center Network infrastructure Design v2.0 (DCNID) Designing Cisco Storage Networking Solutions 4.2.1 (DCSNS) Visit www.globalknowledge.com or call 1-800-COURSES (1-800-268-7737) to speak with a Global Knowledge training advisor.

Copyright 2011 Global Knowledge Training LLC. All rights reserved.

11

About the Author


Ray Dooley is a co-founder and CEO of ICM in Redmond, WA. He has been a Global Knowledge instructor since 1994 and teaches the entire CCNP curriculum, plus advanced classes in MPLS, BGP, ATM, and Cisco Troubleshooting. He has recently become involved in data center and storage network design.

Copyright 2011 Global Knowledge Training LLC. All rights reserved.

12