A few days ago, microprocessor giant Intel (INTC 0.06%) formally unveiled its seventh-generation Core processor family for desktop computers, known as Kaby Lake. With the launch came new chips from Intel as well as new motherboards using the company's new Z200-series motherboard chipsets (the top of the line ones targeted at high-end personal computers are known as Z270).
The main new feature that came with the Z270 chipset is support for the company's new Optane memory technology. Optane is the marketing name for the company's upcoming 3D XPoint memory technology, a new type of memory that's supposed to be much faster than the NAND flash used to build today's solid state storage drives.
3D XPoint memory chips, which Intel markets as Optane. Image source: Intel.
Intel's marketing pitch for Optane is simple: a small Optane-based solid state drive can be installed into a system alongside a traditional hard disk drive to deliver "[solid state drive]-like speed and [hard disk drive] capacity."
Image source: Intel.
This all sounds like a cool idea, and in theory, could boost Intel's PC dollar content share while also helping the hard disk drive makers. Let's look at how it could do so, if the technology works as advertised (a big "if").
The advertised value proposition
The value proposition, particularly for desktop personal computer users, goes something like this. Hard disk drives, which are produced primarily by Seagate (STX +0.00%) and Western Digital (WDC +5.11%), are generally very inexpensive on a per-gigabyte basis. To put this into perspective, it costs about $260 to purchase a 1-terabyte solid state drive for a notebook or desktop computer (one can spend much more for faster/better quality drives, though). A 4-terabyte hard disk drive (that's four times the storage capacity as the solid-state drive mentioned above), runs only about $163 as of a recent check.
This means that the cost per gigabyte (or, more generally, cost per unit of storage capacity) of hard disk drives is still substantially lower than those of solid state drives, even as the latter's cost has come down considerably over the years.
For users who require lots of storage space for, say, high-resolution videos, video games (a single modern triple-A game can take up over 100 gigabytes of space on a computer disk drive), or other large programs, building systems with all flash storage can be prohibitively expensive.
At the same time, solid state drives are much faster than hard disk drives, so users may want to keep performance-sensitive applications on a solid-state drive, while relegating the remaining files and programs to higher-capacity but slower hard disk drives.
The value proposition that Intel and its partners are advertising goes something like this: A small, but very fast Optane-based solid state storage drive (Intel says Optane is substantially faster than traditional NAND flash) acting as a "cache" to a large hard disk drive can allow a system to enjoy similar storage performance to one that's equipped with NAND-flash based storage drives for a much lower overall cost.
Potential winners and losers
If the technology works as claimed and sees significant adoption, then there could be winners and losers.
Among the potential winners, I see Intel, as it would be able to gain content share in personal computers by selling Optane-based solid state drives (Intel does not supply hard disk drives today, and while it does build consumer NAND-flash solid state drives, its share in the consumer market isn't high).
The hard disk drive makers, too, could benefit. If Optane combined with hard disk drives really delivers the performance of all-flash desktop computers for a much lower price, then this could erode the value proposition of all-flash storage setups in some cases, potentially slowing the shift from hard disk drives to solid state drives in desktop personal computers.
The losers would be vendors of solid state drives, as Optane Cache + hard disk drive could reduce demand for traditional flash-based solid state drives.
A big caveat
There have been lots of schemes like what Intel is proposing with its Optane caches done in the past, so this isn't exactly new. Hard disk drive makers have (and in some cases still do) offer hard disk drives with small NAND flash caches built in to try to deliver performance that's close to that of pure solid state drives.
In 2011, Intel's Z68 desktop platform included a technology called Smart Response Technology (SRT) that allowed users to pair small flash-based storage drives with large hard disk drives to try to achieve solid state drive-like speed on the cheap.
Per AnandTech's testing, SRT didn't deliver "the same performance as a large dedicated [solid state drive]" but it could allow users to "turn any hard drive into a much higher performing storage device."
What could potentially make the Optane-based version of this concept more interesting than previous efforts is that instead of pairing a small NAND-flash based solid state drive with a large hard disk drive, Intel is pairing Optane technology (which Intel says is much faster than NAND flash-based drives) with hard disk drives to try to get NAND-flash solid state drive-like performance.
It'll be interesting to see how this technology fares in the real world and, perhaps more importantly, if system builders and consumers ultimately adopt it.
