Within the data center, a debate rages over where and when to apply Flash memory. Many IT organizations are clearly embracing Flash memory first in the form of a card that plugs into their server. But at the same time, Flash memory is finding its way into both all-Flash arrays and hybrid storage systems made up of solid-state drives (SSDs) and magnetic storage.
Violin Memory is betting that the ultimate answer is going to be all of the above. After deciding to focus its efforts on all-Flash arrays, Violin Memory this week announced that it is making synchronous and asynchronous replication and stretch metro cluster capabilities available on its Concerto 7000 Series all-Flash array alongside storage snapshots, thin provisioning, LUN and capacity expansion, advanced data protection and dynamic storage scaling functions.
The Concerto 7000 series, now generally available, can be configured with up to 280TB of solid-state disk (SSDs) in a fully configured 18RU rack that Eric Herzog, chief marketing officer and vice president of business development for Violin Memory, says now pushes the street prices of SSDs in an enterprise IT environment below $4 per gigabyte.
Rather than charge customers for every data storage service, Herzog says Violin Memory is moving to lower the total cost of deploying an all-Flash array by bundling many of the software applications that other vendors charge for within the cost of the array. Violin Memory will not charge for applications that are typically used within a data center environment, but will continue to charge a fee for applications that involve, for example, replicating data between multiple data centers.
Herzog says the storage industry as a whole has reached an economic tipping point where using Flash on both the server and within arrays makes sense. Flash storage on servers is typically available to only a limited number of applications. Herzog says Flash arrays make more sense when IT organizations need to share access to SSDs across a large number of applications.
Ultimately, Flash storage makes sense not just because of the additional performance boost it provides, but because it reduces the complexity associated with providing primary storage access to high-performance applications. Rather than trying to optimize performance by “short stroking” magnetic drives in a way that limits the amount of data that can be written to particular sectors of the magnetic drives, SSD gives applications consistent access to terabytes of data in a way that provides more consistent levels of performance. In contrast, accomplishing the same goal using magnetic storage winds up wasting terabytes of unused space on a magnetic drive that is being “short stroked” to optimize performance.
When you add up all the math, Flash goes well beyond the cost per gigabyte. It’s really about providing the most consistent level of access to high-performance primary storage possible in a way that doesn’t require application developers to master the nuances of spinning disks to achieve it.