Page II: In this issue of Industry Insider, Linus Wong, our guest columnist from Adaptec Storage Solutions, traces the origins of Serial ATA (SATA), a relatively new connectivity interface.
As a replacement for UDMA disk drives in desktop systems, Serial ATA has been a success. The initial promoters of the technology built strong and increasing industry support by maintaining a consistent and well-defined charter. The standards process was private, controlled by five companies, but well managed and open to broad industry participation. Today, Intel's south-bridges include two Serial ATA ports and disk drives are available at multiple capacity and price points. Vendors are delivering on the promise of task file register software compatibility, bringing backward compatibility of Serial ATA controllers to existing ATA drivers already embedded in commercial operating systems.
More desktop platforms are standardising on Serial ATA as the primary disk storage interface and, at this juncture, a full transition from ATA to Serial ATA on the desktop is limited only by the availability of Serial ATA CD and DVD drives. Current CD-ROM and DVD drives - which use existing parallel ATA technology - remain in widespread use.
The original Serial ATA specification was never meant to meet the needs of the high-end server and networked storage markets. To control development costs, Serial ATA 1.0 was developed without enterprise-class capabilities such as an advanced storage management protocol, scalable bandwidth for large numbers of drives, broad command queuing, active dual-port capability, and support for simultaneous multi-initiator access required for clustered environments.
Desktop drives in enterprise environments
As Serial ATA moved from laboratory to production in preparation for desktop deployments, a new class of data was rapidly growing: reference information. More small and mid-size businesses with modest IT budgets were digitising information requiring bulk storage including e-mail, presentations, graphics and a variety of images including CAD/CAM drawings, medical X-rays and bank cheques, increasing demand for low-cost storage and forcing these businesses to seek a cost-effective alternative to high-end disk drives for storage of these infrequently accessed types of data.
At the same time, many companies began to assign a value to information based on age, classifying data into two types: primary and secondary. Primary data -- usually high-demand information used in applications such as credit card, bank and customer service databases, and e-commerce -- requires high-performance transactional disk drives that can withstand the rigours of round-the-clock access. As primary data ages, storage costs begin to outweigh performance, and businesses archive the data to lower cost disk drives in an Information Lifecycle Management (ILM) strategy that lends itself to using high-capacity Serial ATA desktop drives for bulk storage.
Near-line storage applications -- such as disk-to-disk backup -- are another example of a workload where capacity is more important than performance. Serial ATA drives are ideal for both of these cost-sensitive environments since Serial ATA storage is optimised for capacity whereas enterprise-class disk drives such as SCSI are optimised for performance and reliability. Disk-to-disk backup is a growing practice as more businesses use the Internet to support round-the-clock operations and need to stage data for backup on disk drives before moving it to tape storage to maintain uninterrupted business operations.
Desktop drives, because they are designed for the light duty cycle associated with normal business hours versus the 24x7 demands of enterprise storage, raise reliability concerns among IT managers wanting to deploy Serial ATA in the enterprise. High-port count Serial ATA RAID controllers have emerged to allay data availability fears associated with deploying numerous desktop drives in these mission-critical storage environments. RAID technology provides uninterrupted data access in the event of a disk drive failure.
Looking ahead
With the completion of the Serial ATA 1.0 specification, the industry has turned to beefing up the protocol with enterprise enhancements. The goal is to provide a low-cost scalable interface that meets the infrastructure needs of high-performance transactional environments. Known as Serial ATA II Extensions, the updated specification defines features such as native command queuing, a 3-gigabit-per-second data transfer rate, basic enclosure services, staggered spin-up of disk drives, and new devices such as port multipliers to improve scalability and port selectors to improve reliability.
14%
7%
I think a lot of the problem with the uptake is that it (currently) provides little benefit. It is no secret that many of the SATA drives simply have a PATA->SATA convertor on the drive itself. At the motherboard side, it is a similar story.
Many of the drives were designed with the restrictions that come with PATA. Changing the interface and cable may give the device more potential to be quicker, but many existing devices are throttled so they can work on either.
In short, we have the option of paying more money for no noticable improvement in speed. The seek time and MB throughput (the figures that actually make the thing faster or slower) are identical.
Can't imagine why it hasn't taken the world by storm ...