In blog last year, I had a bit of a pop at SSDs following a failure. I do not as a rule draw general conclusions from specific cases so I didn't assume that SSDs were bad, although there's certainly a warning to be gleaned from the sudden failure I experienced.
I do though feel that a more dispassionate view of the SSD and its impact on storage architectures and designs is worth airing, and it's this: are high-end hard disks about to disappear?
Solid-state disks used to be highly specialised pieces of equipment, but the price per GB is reducing in line with Moore's law, as a result of which there are increasing numbers of products incorporating the technology, and a growing number of ways of using its capabilities. If you doubt it, check the amount of venture capitalist cash that's pouring into SSD-based equipment vendors. For example, IO Turbine, Kaminario, SolidFire, and Virident Systems have all raised millions of dollars, while Fusion-io pulled off an IPO.
We can now find SSD technology deployed in many ways. It's being installed inside the server as super-fast, direct-attached storage, as a tier zero in storage appliances connected over the SAN or Ethernet, and in direct-attached, SSD-only lumps of storage, for example.
Each approach has its own attractions and dis-benefits, with those relating to the connection type being no different from those of conventional storage. The key attractions of SSDs from a datacentre perspective are of course performance, power and space.
Performance is probably the key issue though. For a virtualised environment, which majors in random rather than sequential access requests, getting enough IOPS from spinning disks is a struggle. You can end up buying far more capacity than you need, even if a lot of it is locked out by disk vendors using short-stroking as a way of boosting access times, just to get enough throughput. But more disks means more power drain and more space needed to house them and their cooling requirements.
SSDs ameliorate many of these problems. They don't go away -- what problem ever does? -- but they give you more options by relieving the datacentre of those bottlenecks.
So why would still buy fast spinning storage? It's still a lot cheaper per GB of course and SSDs do still, I feel, need to fully prove themselves in real world usage. And if you buy into SSDs, you still need a good answer when the CFO comes round and asks why so much is being spent on so little storage capacity, especially given that technology's well-known propensity to wear out in relatively short order.
Some of that answer involves of course SSDs' savings on space, power and cooling, but better performance to permit, for example, desktop virtualisation, with its high levels of random IO but also potentially big savings, could be the killer. And there are more products available and coming along that mix the two technologies and promise to provide the best of both worlds; for example, Nimble Storage's CS-Series arrays fall into this category.
All this leaves high-speed enterprise-level HDDs in a bit of limbo. Obviously, they're not going away tomorrow, this year, or even next. But will enterprises be buying them for their primary storage in five years? Or will slower but much cheaper SATA drives displace them for back-end and archiving purposes? What are your buying plans?