It used to be that you basically had a choice of a 5400 RPM hard drive or a 7200 RPM hard drive.  5400 RPM was slow, and 7200 RPM was, well, if not exactly fast, then at least less slow.

Then SSDs came along and took over the high end of storage performance.  SSDs are expensive, however, so if you can't afford one, then a 7200 RPM hard drive was a worthwhile boost in performance over a 5400 RPM drive for not much added cost.

Now we have news that one of the two major hard drive manufacturers is abandoning 7200 RPM hard drives.

http://www.xbitlabs.com/news/storage/display/20130228212108_Seagate_to_Cease_Production_of_7200rpm_Mobile_Hard_Drives_This_Year.html

What makes this odd is that in desktops, Seagate recently went in the opposite direction, ending production of slower hard drives in order to go 7200 RPM only.  Or rather, 7200 RPM for consumer drives; they still make 10k and 15k enterprise drives.

Seagate hopes that the upcoming "Seagate Laptop SSHD" (to replace the Momentus XT brand name) will be able to fill in as a high end laptop hard drive.  The problem is that there's nothing high end about it.  8 GB of NAND flash as a read-only cache means that if you want to write anything, or want to read anything that isn't in the cache, you're stuck with a glacially slow 5400 RPM hard drive.  That is, most of the time, it's a 5400 RPM hard drive.

Hybrid drives may well make some sense.  But hybrid drives with only 8 GB of cache most certainly do not.  The problem is that you have a lot more than 8 GB in data that you regularly read.  Windows alone takes more space than that.  So do a lot of games.  On top of that, sometimes you have to write to storage, too.  If you rely on a hybrid drive with only 8 GB of cache, then you're mostly relying on a 5400 RPM hard drive.  You'd be better off with a simple non-hybrid 7200 RPM hard drive.  That would give you better performance with fewer things that can go wrong--and it would be cheaper, too.

To make matters worse, if you defragment a Momentus XT, you lose everything that is in cache, which kills your performance.  If you don't defragment it, you end up with a badly fragmented drive that you frequently have to read from, which also kills your performance.  If Seagate can't do better than that, then a 7200 RPM hard drive is clearly the better option.

How much cache would it take to make a hybrid drive that is actually good?  I'm guessing that somewhere in the ballpark of 64 GB would get it done.  The problem is that 64 GB of NAND flash plus a controller is expensive enough that it easily puts you into competition with real SSDs.  Those prices should come down with subsequent die shrinks--but losing flash endurance with such die shrinks may make future NAND flash unsuitable for use as small amounts of cache.

The hard drive manufacturers have all seen the writing on the wall in competition from SSDs diminishing demand for hard drives, and especially for lucrative higher end hard drives.  Samsung and Hitachi responded by getting out of the hard drive market entirely.  Seagate and Western Digital have tried making SSDs, and failed miserably at it.  Seagate's hybrid attempts haven't produced worthwhile products, either, though without seeing sales figures, I don't know how successful they have been at tricking people into believing that a hybrid drive is like an SSD in any ways that matter to consumers, other than the higher price tag.

Your post jogged my memory about an article I read from a few years ago about an IBM flash memory breakthrough. I wonder why we haven't seen any products yet?

"Made in IBM Labs: IBM Scientists Demonstrate Computer Memory Breakthrough

Reliable multi-bit phase-change memory technology demonstrated

Zurich, Switzerland - 30 Jun 2011: For the first time, scientists at IBM Research (NYSE: IBM) have demonstrated that a relatively new memory technology, known as phase-change memory (PCM), can reliably store multiple data bits per cell over extended periods of time. This significant improvement advances the development of low-cost, faster and more durable memory applications for consumer devices, including mobile phones and cloud storage, as well as high-performance applications, such as enterprise data storage.

With a combination of speed, endurance, non-volatility and density, PCM can enable a paradigm shift for enterprise IT and storage systems within the next five years. Scientists have long been searching for a universal, non-volatile memory technology with far superior performance than flash – today's most ubiquitous non-volatile memory technology. The benefits of such a memory technology would allow computers and servers to boot instantaneously and significantly enhance the overall performance of IT systems. A promising contender is PCM that can write and retrieve data 100 times faster than flash, enable high storage capacities and not lose data when the power is turned off.  Unlike flash, PCM is also very durable and can endure at least 10 million write cycles, compared to current enterprise-class flash at 30,000 cycles or consumer-class flash at 3,000 cycles.  While 3,000 cycles will out live many consumer devices, 30,000 cycles are orders of magnitude too low to be suitable for enterprise applications. (see chart for comparisons)." http://www-03.ibm.com/press/us/en/pressrelease/34949.wss

Originally posted by Sal1

Your post jogged my memory about an article I read from a few years ago about an IBM flash memory breakthrough. I wonder why we haven't seen any products yet?

It's one thing to demonstrate that you can make one cell that works.  It's quite another to demonstrate that you can make chips with 17179869184 cells each that work over 99.9999999999999% of the time, can withstand many reads and writes without wearing out, can read, write, and erase fast enough to be useful, and you can make them cheaply enough to sell them profitably.  The latter is what you need to do if you want it to be used in consumer SSDs.

And no, those aren't just random numbers that I made up.

It's entirely possible that phase change memory could be used in low-volume enterprise products long before it comes to consumer products.  For some enterprise uses, if you absolutely have to have something that has particular properties that NAND flash lacks, a very high price tag may be acceptable.

Remember that NAND flash was invented in 1980, but didn't appear in consumer SSDs until 2007.