January 10, 2009
SCSI stands for Small Computer System Interface and is often called Parallel SCSI. The technology is
over 25 years old and has a hard time keeping up with today's heavy demands for intensive hard disk reads
Its counterpart, SAS, stands for Serial Attached SCSI and was developed to address the important limitations
attributed to Parallel SCSI. This brief highlights the differences between the two hard drive controller
interfaces, and helps to point out the many features that explain the increasing popularity of SAS.
SCSI is simply a small set of industry standards for physically connecting computers, servers and
peripheral devices together and then transferring aome data between them. Such industry standards help define
overall system commands, the various protocols, the electrical and finally, they describe the optical
interfaces that are in use.
Overall, SCSI is most commonly used for hard disks and tape drives, but it can also be used for a range of
other peripheral devices as well, such as scanners, CDs, DVDs and of course DVD burners.
Often called parallel SCSI, this technology is based on the information bus principle. But being so old,
parallel SCSI maxes out at a speed of about 320 MB/sec and its performance gets degraded even more as additional
devices and various peripherals are added to the shared bus, neither of which can be compromised due to today's
computational complexities in the prevailing corporate IT requirements.
For its part, SAS (Serial Attached SCSI) was developed to address the numerous I/O (input and output)
and direct-attach disk storage requirements that traditional parallel SCSI technology simply couldn't meet.
On one hand, the newer technology offers logical compatibility with SCSI. It also provides the reliability,
performance, scalability and manageability that IT professionals have come to expect of SCSI. Like parallel
SCSI, SAS is a data-transfer technology specifically designed to move data to and from computer storage
devices such as hard drives and tape drives.
Unlike SCSI, which is multi-drop, SAS is a point-to-point protocol and allows for much higher speed data
transfers than has been possible with parallel SCSI.
It also uses the standard SCSI command set for interacting with SAS End devices.
The SAS protocol was developed and is maintained by the so-called "T-10 technical committee" of the
International Committee for Information Technology Standards (INCITS).
A typical SAS system consists of the following four essential components:
The service delivery subsystem
An Initiator is a device that sends service and task management requests to a target device and
receives data responses for the same requests from the target device. Initiators may be provided as an
on-board component on the motherboard or as an add-on host bus adapter.
A Target is a device that contains the logical units and target ports. It receives device service and
task management requests for processing and sends responses back to the initiator that transmitted these requests.
A target device could be a hard disk or a disk array system.
A SDS (Service Delivery Subsystem) is part of the I/O (input/output) system. It transmits the information
going back and forth between an initiator and the target. Generally, an SDS consists of the cables that
connect an initiator to the target with or without expanders.
Overall, expanders are the devices that are part of an SDS. They facilitate communication between the SAS
devices. They also facilitate connection of multiple SAS devices to a single initiator port.
From the end-user point of view, SAS provides enterprise-class robustness, protection of investments in
compatible SCSI software and applications, and, since it is compatible with SATA (Serial Advanced Technology
Attachment), SAS offers a better choice of direct and attached data storage devices in a single SAS system.
Since it's based on a serial interface, SAS also allows for increased device support as well, which
is an added bonus.
Source: IBM IT Services.