|
|
IDC: SSD performance gap 'small'
Author: Gareth HalfacreePublished: 17th July 2008
If you were planning on an SSD as a drop-in replacement for a 7,200 RPM disc, don't expect a performance boost.
The report, quoted by SearchStorage yesterday, suggests that tests showing a high performance differential between SSD devices and traditional mechanical drives might be a trifle misleading – and IDC's own testing bears that out.
The main complaint the firm has with normal testing is that the comparisons are always between the latest and greatest SSD devices and 4,200 RPM mechanical drives. While the comparison makes sense from the consumer point of view – laptops offered with SSDs usually ship with a 4,200 RPM drive if you choose to save your dosh – they don't make much sense in a world where 5,400 RPM mechanical drives are the norm, and top-end laptops are shipped with 7,200 RPM drives.
IDC researcher David Reinsel compared 7,200 RPM drives with SSD drives and, perhaps most importantly, took the performance of the entire system into account when compiling his results. The conclusion of the research is that previous lab tests on isolated devices may have been misleading, and that the performance gap between the two technologies – while definitely present – is nowhere near as dramatic as SSD proponents have been making out.
Reisnel also point out that the difference in the architecture of SSDs and mechanical drives mean that, despite physical compatability, the devices aren't a simple drop-in replacement at the enterprise end of the market. Rather, many enterprise systems will require a complete redesign in order to achieve the best possible performance from solid-state storage devices. Reisnel is quoted as saying that “there will be what's called a 'period of interdependency' with this technology,” and that the solution won't be simply “plug and play.”
The results of the test are only available to subscribers to IDC's white paper service, so actual figures are currently hard to come by. The company is also being secretive about actual testing methodology, with no mention of what software was used to measure performance on the test units. Nevertheless, if true it is perhaps another blow for the SSD market – currently trying to win back support after researchers discovered that the power saving potential of the vast majority of devices currently on the market was a lot lower than anticipated.
Is the research enough to make you think twice about that expensive SSD investment, or do you need hard figures to be shared before you'll decide one way or the other? Share your thoughts over in the forums.
Related articles on bit-tech.net
Comments (8)
Stats: 0.096 seconds
thats a difference of 2.99eur per gigabyte - or a 2,400% premium. still sound like a viable option to you? not me.
SSD has a future for much greater speed. Intel has said their SSD is going to be 200+ MB/s ... although it still hasn't appeared yet.
putting Vista on Raptor greatly increases Windows smoothness, so putting on a high performance SSD should increase it even more.
the "small gap" is just saying in comparison to high speed HDD, it's not a large improvement. after-all, current high speed HDD (veloci-Raptor) is already very fast
Now, you can turn the pagefile off and reduce the number of temporary files created (or move both to a mechanical volume), but that's something you have to do. Likewise, you can do the same in Linux and, I assume, MacOS. That doesn't make either of those 'optimised' for SSD use either - unless someone has already done so and released a special version, such as the SSD-optimised Eeebuntu Ubuntu variant designed for the Asus Eee PC.
If a controller chip burns out, can you still recover your data? Is there such a thing as a block-by-block clone of the SSD? How do software tools such as Data Rescue II cope with SSDs?
The wear levelling algorithms will ensure that individual sectors do not get prematurely destroyed by pagefiles etc.