There’s clearly a massive amount of interest in Flash storage these days as a way to alleviate performance bottlenecks. But one of the challenges IT organizations are likely to quickly encounter is trying to manage Flash-based storage systems at scale.
According to Kaminario CEO Dani Golan, the major storage vendors have all built monolithic storage architectures around controllers that not only don’t scale as cost effectively, they extend proprietary systems that lock customers into that architecture.
In contrast, Golan says Kaminario K2 Flash storage systems are built around a scale-out that allows multiple Flash arrays to appear as one single logical pool of storage. Golan says that approach allows IT organizations to not only add Flash storage over time in the most non-disruptive way possible, it’s ideal for Big Data environments where the location of different data sets is likely to be distributed across multiple Flash arrays.
While the cost per gigabyte of Flash storage is still higher than magnetic disk drives, the fact that Flash storage systems improve performance in a way that doesn’t require developers to optimize where data is placed on high-performance magnetic disk drives means that Flash drives introduce a lot more agility into the overall IT equation. Flash drives clustered in a pool that is essentially an on-demand service also take a lot of vagaries out of the I/O performance equation, which is one reason why, in the event of a drive failure, Kaminario now guarantees performance will never drop more than 25 percent across a K2 array.
The fact that Flash storage arrays also consume a lot less power and space means the total cost of ownership associated with Flash storage starts to look pretty compelling.
There is, of course, no shortage of Flash storage options these days so IT organizations will probably wind up deploying a mix of all Flash, hybrid and magnetic disk storage systems over time. But the one thing that is for certain is that the economics of storage in the data center are fundamentally changing in a way that IT organizations can no longer afford to ignore.