By Scott M. Fulton, III, Betanews
The solid-state disk drive is supposed to be fast. After all, it's mostly made of memory -- and last we checked, flash RAM was fast. In practice, however, some applications with SSDs can be slower than with HDDs, the reason being the way data is cached as it's collected and moved through I/O channels into system RAM.
The transfer interface is the bottleneck, and the engineers that contribute to the Serial ATA (SATA) transfer specification admit that fact openly. Just a few years ago, you might never have thought that 3 gigabits per second (Gbps) would end up causing problems; but as it turned out, the faster SATA 2.0 maximum transfer rate enabled new applications, which ended up introducing users to those bottlenecks for the first time.
Now, the SATA-IO organization is preparing to eliminate that logjam, with the publication this morning of the SATA 3.0 specification. Its goal is to accelerate maximum transfer speeds to 6 Gbps, and in so doing, widen the bandwidth between components where these new bottlenecks have recently been introduced.
"SSDs provide faster data access and are more robust and reliable than standard HDDs because they do not incur the latency associated with rotating media," states a recent SATA-IO white paper. "SSDs are used in a variety of applications but one of the most exciting is two-tier, hybrid drive systems for PCs. The SSD serves as short-term and immediate storage, leveraging its lower latency to speed boot time and disk heap access while a HDD, with its lower cost per megabyte, provides efficient long-term storage. With SATA 3 Gb/s, SSDs are already approaching the performance wall with sustained throughput rates of 250-260 MB/s [megabytes per second, note the capital "B"]. Next-generation SSDs are expected to require 6 Gb/s connectivity to allow networks to take full advantage of the higher levels of throughput these devices can achieve."
The rapidly improved transfer rate may also increase not only the efficiency but also the lifespan of conventional hard drives, by introducing a concept called native command queueing (NCQ). With data transfer and data processing threads operating at roughly parallel speeds today, the only way existing HDD controller cards can synchronize these processes is by running them in sequence. That eliminates the opportunity controllers might have to read some data out of sequence (similar to the way Internet packets are received out of sequence) and assemble them later. By doubling the theoretical maximum throughput rate, HDDs can read more data from rotating cylinders along parallel tracks, without having to move the head...and that reduces wear on the drive head. Of course, a new generation of drive controllers will need to be created to take advantage of this capability.
However, the best news of all is that a new generation of SATA cables does not have to be created. We can all use the cables we have now, to take advantage of the performance gains to come. Though the new controller cards will themselves be replacements, the SATA interface itself doesn't change to the extent that new cables are required. So existing external equipment, including the latest models in the exploding realm of external HDD storage, will be fully compatible.
SATA-IO is making no promises as to how soon consumers will start seeing the new controllers, or PCs where those controllers are pre-installed.
Copyright Betanews, Inc. 2009