Intel has announced two new solid state disk drives made on its leading-edge 34nm process. The two new SSDs are X25M SATA parts weighing in at 80GB and 160GB, and they're meant to replace Intel's existing X25M drives in those capacities, but at 60 percent less cost and with better performance. The 80GB X25-M is $225 in lots of 1,000 (down from $595), and the 160GB is $440 (from $945). That's some serious discounting, and it may well drive even more SSD uptake in the coming quarters despite the ongoing IT spending crunch.
So what do you get for 60 percent less? In a word, speed. The new drives boast a 25 percent reduction in read latency, which was already about 60x the speed of an average hard disk; write performance has also doubled with this new generation.
Incidentally, the yawning gap between SSD performance and HDD performance is behind the push to SSD-specific interfaces on I/O hubs now, like with Intel's upcoming Pine Trail platform. The bottlenecks are really at the system level, and possibly in the OS, as well. Aside from some halfway steps like Intel Turbo Memory, there hasn't really been a major rethinking of PC system architecture that starts with the presumption of a very fast, large pool of nonvolatile storage (e.g., SSD) that sits behind main memory. There may have to be even more SSD adoption before such a rethink can become feasible, though.
Noticeably absent from the announcement is a high-capacity 320GB drive, which many were expecting. So why did Intel choose to hold off on pushing the SSD capacity envelope and instead produce drives that are direct replacements for existing 50nm products? Like most things nowadays, the answer has to do with the economy.
A 320GB drive would be an expensive, premium product of the sort that nobody is buying right now—not tapped-out consumers, and not cash-strapped IT departments. So instead of trying to sell a high-end SSD, Intel took advantage of Moore's Law to massively cut the cost of making its existing flash products, so that it could lower the price and sell more of them into a price-sensitive market. Intel could even be making a bit more profit on each drive even while passing considerable savings on to customers.
It's also the case that the 34nm transition was in the works well before the financial crisis, and would have been in the latter stages of coming online when Intel's revenues fell off a cliff in the last quarter of 2008. So Intel has gone ahead and brought that capacity online, and they're going to use it to make what everyone wants right now, i.e., cheaper storage that, as an added bonus, also performs a bit better.
0 Comments