Violin Memory, a pioneer in all-flash arrays (AFA) that fell on hard times after their IPO, announced today that they have been acquired by Quantum Partners LP, a private investment fund managed by Soros Fund Management LLC.
Soros acquired Violin's assets in a bankruptcy auction, exchanging $26 million in unsecured debt for $15 million in equity in the reorganized company.
The Storage Bits take
Violin took the hard road to building an AFA. They focused on a hardware solution that, in my judgement, was an innovative platform. In TPC-C benchmarks the Violin array offered very low and consistent latency (see The SSD write cliff in real life on StorageMojo).
In a world where every 3U AFA offers a half million IOPS or more, the key to maximum performance is latency, not IOPS. But designing hardware isn't the fastest path to market.
Other AFA vendors focused on using commodity SSDs with dual RAID controllers and plenty of software features that customers expected from their existing disk-based arrays. Since those controllers couldn't see when the SSDs would decide to perform garbage collection and other functions, those arrays had higher and much more variable latency.
But the IOPS wowed customers, the array features fit with their existing management processes, and the latency issues didn't come to the fore until later. Now most AFA vendors are moving away from standard SSDs to their own hardware flash blades, similar to what Violin did five years ago.
Sometimes innovation can be too much of a good thing. It will be interesting to see if the Soros capital injection will reinvigorate this early AFA leader.
Courteous comments welcome, of course. I've been a fan of the Violin architecture since I was first exposed to it eight years ago. Since then I've done work for Violin. Bottom line: I'm biased and am glad to see that their IP is getting another shot.