Violin Memory Gives IT Control over Flash Array

Software-Defined Storage: Driving a New Era of the Cloud

Violin Memory wants to give IT organizations more control over when and how inline data deduplication and compression are applied on Flash memory arrays. This week, Violin Memory announced the Concerto 2200 series Flash Memory array that can deliver up to 672TB of usable storage when data deduplication is invoked.

But Eric Herzog, chief marketing officer and senior vice president of alliances for Violin Memory, says that what distinguishes the Concerto 2200 series most is that Violin Memory doesn’t force IT organizations to turn on data deduplication and compression. While having those capabilities makes sense for some classes of applications, Herzog says there are instances where having those functions on a Flash array are either simply redundant or they adversely affect application performance.

Rather than force IT organizations to always turn on those features, Herzog says the Concerto 2200 Flash arrays give IT organizations the option to turn those features on or off depending on the requirements of the workloads to be supported.

The Concerto 2200 includes a dashboard that enables IT organizations to see the actual data reduction rates being achieved. The array itself consists of two appliances that can not only deduplicate and compress data, they can also share groups or files from NFS and write the data to LUNs on either a Violin 6000 or 7000 series All Flash Array. The LUNs being serviced can all be on one array, or scaled across up to four arrays. Given those capabilities, Violin Memory claims the Concerto 2200 solution is capable of storing enough data for 2,500 persistent desktops in a VDI deployment at a street cost of approximately $1.81 per GB or $75 per desktop.

It’s still unclear to what degree Flash arrays will be adopted across the enterprise. But given the diversity of the workloads in the enterprise that need access to as much I/O as possible, it’s clear that Flash memory is going to show up both on the server and in arrays, which more easily provide access to Flash memory resources and can be shared by a distributed set of servers running multiple applications.