Only way to stop the iPad: Flash-disk mutant SPEED FREAKS
At least... until hybrids inhale helium
Money: It's the BOM
NAND costs money and disk drive buyers, especially OEMs, want to buy drives at as low a cost as possible. The Bill of Material (BOM) cost has to meet this need. Adding dollars with NAND caches needs justifying; the vendors have to be convinced there is a market. That encourages the disk drive manufacturers to be parsimonious with the amount of flash they add. There has to be as big a difference between the cost/GB of all-flash storage and hybrid flash/disk storage as possible, and the smaller the amount of flash in the hybrid drive the greater this difference is.
Hence Seagate's Momentus XT has 0.8 to 1 per cent of its disk capacity duplicated in flash. That means, for example, Excel and Word app load times will be at flash speed but spreadsheet and Word documents will be loaded from disk and saved to disk at disk speeds. The bigger these files get, the slower their access will be be. But them's the rules: hybrid drives are a halfway house between all-flash and all-disk - you get some flash speed added to your disk capacity for a little more than all-disk money.
Content consumption vs creation
This won't matter so much for tablets where data consumption takes precedence over content creation. It will be a benefit to be able to store many, many more videos and photos and music tracks for your money than you would with an all-flash tablet. Compare, for example, a 124GB flash tablet with a 512GB hybrid flask/disk tablet and you get three times more space for content with the hybrid.
The ratio of content creation to consumption will incline more to creation as we change from table to notebook, whether ultrathin or not, and again as we change from notebook to desktop. The way desktop designs are changing towards all-in-one screen and system units means a desktop these days is practically a notebook mounted vertically behind a large screen, albeit a notebook with more real estate inside the thin enclosure. This trend encourages the use of flash caching to provide speed instead of having two disk drives, where that was the case.
Extrapolating into the future
Two trends appear likely to change the coming hybrid drive scenarios we've painted. One is three-level-cell flash (TLC), and the other is helium gas-filled drives.
Current hybrid drives use fast and expensive 1-bit or single level cell (SLC) flash. All-flash drives typically use less expensive 2-bit or multi-level cell (MLC) flash. As it stores two bit values per cell, it holds twice as much data in the same space as SLC flash at the cost of slower access speed and a shorter working life. TLC flash has a much shorter working life and a slower access speed but costs less than MLC flash. Controller technology, like STEC's CellCare, can extend TLC endurance to acceptable levels meaning that an all-TLC flash drive will be much more affordable than an all-MLC flash drive.
The price premium you have to pay for a TLC flash tablet or notebook will be significantly less than for an MLC one, making the price difference with a hybrid tablet or notebook less and so prompting users to pay the extra money needed to get all-flash speed.
We don't see this happening with desktops though. Helium-filled drives, with helium gas offering less resistance than air to spinning disk platters and moving heads, enables a 3.5-inch drive to have more platters in the same space, six or seven instead of four or five. This means today's 4TB drive could be tomorrow's 6TB drive using the same recording technology level - or the 8TB drive which emerges later in the future using an even more advanced recording technology.
The cost/GB differential between an all-flash drive and a hybrid, helium-filled drive will remain wide, even if the comparison is with a TLC flash drive, and so, we believe, hybrid disk drives will persist in desktop PCs, evolving to hybrid helium-filled drives.
This pattern may occur with notebook computers too if they use multi-platter 2.5-inch drives. Ultrathin notebooks will probably be constrained to single platter drives because of dimensional constraints. Thicker notebooks could use 2- or 3-platter drives and these could be helium-filled drives cramming 3 platters into today's 2-platter form factor, and so increasing capacity by 50 per cent. That could preserve a meaningful cost/GB difference between all-flash notebooks and hybrid flash/disk notebooks and so preserve the notebook 2-5-inch disk drive market which will be threatened by all-TLC flash drives, assuming they are built.
Hybrid disk drives look to have a viable role, certainly in the 2.5-inch form factor, and probably in the 3.5-inch form factor. It appears to us that desktop hybrid helium-filled 3.5-inch drives can survive a transition from MLC to TLC flash drives but TLC drives will tend to eject 2.5-inch disk drives, hybrid or not, from tablets and ultrathin notebooks, not so from thicker notebooks where multi-platter 2.5-inch drives can be used.
We could draw a parallel with cars. A few years ago owners could choose to buy a car for comfort and convenience but forego 4-wheel drive off-road capability and also van capacity levels. But off-road vehicles were clunky and complex and vans were simply not comfortable or well-equipped.
Now we have sports-utility vehicles and people carriers that marry saloon car comfort and convenience with some 4-wheel drive off-road capability on the one hand or with luxurious well-equipped van capacity on the other. Such vehicles are very popular. On that basis hybrid disk drives could also become very popular indeed. ®