SPONSORED: How SAS Can Save Your Data
SPONSORED: How SAS Can Save Your DataSerial-Attached SCSI (SAS) drives are rapidly taking over in a wide array of business applications.
Reducing hard drive types to a good/better/best scenario is too simplistic. A glance at the core ideas of tiered storage might leave you thinking, “I use the fastest drives at the top for high-I/O, transactional tasks; mid-level drives for nearline work; and the slowest enterprise drives for bulk storage.” There are elements of truth in this view, but it can lead to confusion. For example, is it better to have a Serial-Attached SCSI (SAS) hard drive or a Serial ATA solid state drive in a mid- to low-level tier? It turns out there’s more to the question than meets the eye.
In large enterprises, SAS has been gaining ground for several years. But in smaller organizations, where price sensitivity can be even higher, SAS tends to be less well understood. Without question, SAS is not a suitable interface for every occasion. However, a better understanding of what SAS is and how it compares to SATA can help IT purchasers make better decisions and ensure that the right technology is being adopted for optimal total value and return on investment.
SAS Basics
For many years, enterprise storage relied on devices built around the Small Computer System Interface (SCSI) command set. SCSI defined communications between hosts and peripherals, identifying SCSI initiators at one end and SCSI targets at the other. Up to 16 devices could connect to a single SCSI bus, which was a massive improvement over the two devices (one master, one slave) we all used to endure with Parallel ATA drive technology.

The SCSI command set, with all of its programmability and robustness, was great; the parallel interface used with traditional SCSI ultimately was not. Just as USB rose up to overthrow parallel peripheral connections, so too did serial storage interfaces ultimately triumph. The Fibre Channel standard arrived in 1995, carrying mostly SCSI commands. iSCSI, which runs SCSI commands over Ethernet, is now eating away at Fibre’s dominance as a storage network technology. Not least of all, SAS runs SCSI over tweaked SATA cabling. In fact, SAS also runs ATA protocol (as well as SMP), SAS controllers are backward compatible with second-generation (3 Gb/s) or later SATA drives. This is a very important point to keep in mind for later.
SAS started its transition into 6 Gb/s signaling in 2009 while SATA is only now making the same shift. A doubling to 12 Gb/s is on the road map for SAS. We now have SAS drives in both hard disk and solid state varieties, with both MLC and SLC options for SSD and hard drive offerings in 7200, 10K, and 15K RPM varieties. There are very few 10K options in the SATA world and none for 15K.

There are some other notable differences between SAS and SATA. The SCSI command set is better suited to error recovery than ATA’s SMART commands, which obviously appeals to business users. Higher signaling voltages in SAS allow data cables of up to 33 feet whereas SATA can only manage 3 feet (or 6.6 feet for eSATA). Perhaps even more important, SATA controllers identify drives according to which port they’re plugged into whereas every SAS device carries a unique World Wide Name (WWN) identification code. The WWN is key in allowing SAS to function as a large-scale storage network medium. SATA does not enjoy this flexibility and thus scalability.