Brian's Brain

One of the many 'dirty little secrets' of the tech industry is that whereas PMR (perpendicular magnetic recording) bit-packing techniques are improving HDDs' sustained sequential transfer rates, those performance metrics still don't come anywhere close to swamping first-generation 1.5 Gbps SATA's capabilities (far from those of the 3 Gbps or upcoming 6 Gbps descendants). Granted, you can burst data into or out of a drive's integrated cache faster than you can transfer it to or from the rotating magnetic media sitting behind it, but RAM buffers are only tangibly meaningful in certain access profile situations, and they're to some degree mitigated by the system-side buffers supported by modern operating systems.

It's a reality that HDD suppliers (who, like their peers in other areas of the tech economy, have long sold customers on a 'bigger numbers is better' simplistic marketing mantra) will admit begrudgingly, anonymously, and sometimes only when under the influence of lip-lubricating intoxicants. And it's a reality, judging from a recent kerfuffle, that Apple also knows well. When early adopters compared specs of their latest-generation unibody 'Pro' MacBooks, they discovered that whereas HDD-based units reported 1.5 Gbps SATA capabilities in System Profiler (OS X's analogy to Windows' Device Manager), SSD-housing laptops claimed 3 Gbps speeds. Since the Nvidia core logic chipset was the same in both cases, the bewilderment was understandable.

Apple quickly released a firmware update that enabled 3 Gbps potential in all systems, regardless of their factory-installed mass storage subsystems' capabilities. But the release notes contained the money quote:

While this update allows drives to use transfer rates greater than 1.5Gbit/sec, Apple has not qualified or offered these drives for Mac notebooks and their use is unsupported.

Translation: regardless of what System Profiler might have claimed in the past, Apple's to date never shipped a system (even a high-end 'Pro' model) containing a HDD capable of 3 Gbps SATA transfer rates. This decision probably saved Apple a few pennies-to-dollars per system versus using drives that suppliers had qualified to full 3 Gbps SATA speeds, and it's a pragmatic one for the reasons described above.

Am I saying that SATA is of no benefit? Heavens no! Anyone who's ever built up a PC can attest to the simplification superiority of a slim serial interconnect versus the fat, flat parallel ribbon cables of the PATA (i.e. IDE) past. SATA's direct core logic controller-to-mass storage device 1:1 interface is also a notable improvement (performance- and other-wise) versus IDE's master/slave shared bus arrangement. And SCSI-reminiscent NCQ (native command queuing) further accelerates SATA's speed potential, by enabling the HDD to dynamically re-order incoming access requests dependent on what it's doing at the time and where it's read/write head(s) are currently located.

And am I saying that 3 and upcoming 6 Gbps SATA flavors are nothing but marketing hype? Again, no, but my enthusiasm here is specifically devoted to SSDs. My recent hands-on testing showcased SSDs' performance superiority over HDDs, particularly in read-intensive usage scenarios and double-particularly when access profiles contain a high percentage of random fetches. As SSDs' price tags at user-meaningful capacities continue to drop, nearing (and at some point plunging below) the $/GByte metrics of similar-sized HDD counterparts, SSD adoption will continue to expand, and latest-generation SATA versions will correspondingly shine.