It has emerged that both Violin Memory Inc and Pure Storage Inc are on the verge of initial public offerings, based on news articles last week. Chris Mellor reports that Violin have filed for IPO, looking to raise $172m and in separate news, Pure Storage have raised a massive $150m in preparation for IPO some time soon. Although both of these startups are in the same part of the storage industry, notably all-flash storage arrays, I think their target markets and potential IPO outcomes couldn’t be different.
I’ve been tracking Violin Memory for some time and have been fortunate enough to receive regular briefings from the company. I think their technology is solid. The idea of using dedicated VIMMs (like a DIMM card for PC memory, but using NAND) makes sense. It allows the storage protocol to be removed from the internal architecture and so drive higher performance out of their storage. It also enables better workload distribution and removes the need to deal with the specific issues of solid state drives. So if you want a high performing solid SSD array with low latency, Violin is a good choice.
However, performance comes at a price (developing your own VIMMs isn’t cheap) and because of that, their market becomes limited to specific applications and use cases where only a more expensive solution will do. So immediately their market share is limited, which of course directly limits growth and revenue.
So what are they proposing? Taken from Chris’ article, from the S-1 filing:
- Future profitability depends upon sales of its Velocity PCIe flash card product, and hyperscale customers will be pursued for this gear
- It expects to “opportunistically make strategic investments in, or pursue acquisitions of, companies with innovative technologies”.
Unfortunately the PCIe flash card market is pretty saturated already and the perceived market leader, Fusion-IO isn’t doing that well itself. As for acquiring companies, that $172m won’t go far. What other options are there? Here are a few thoughts:
- “Do a Nutanix” – add server functionality to the storage array. This one could be tough though as VMware have just entered this market (albeit with a software only solution) and there is a lot of maturity out there.
- Merge with a competitor – crazy as it may seem, a merger with Fusion-IO could be a way forward.
- Leverage acquisitions – Violin purchased Gridiron Systems earlier this year (January 2013), but I haven’t heard much about it.
I first saw Pure Storage at Tech Field Day 8, then again a year later at Storage Field Day 1. Their technology is based on using “commodity” solid state drives (SSDs) with clever software to drive a consistent sub-millisecond response for I/O, regardless of the workload. There were some fascinating discussions on how this is achieved that basically come down to understanding how SSDs function in certain circumstances (for example once they are “full” and have to perform garbage collection) and working around that. Pure are targeting a different market to Violin, aiming to deliver SSD performance to as wide a market as possible, including virtual server (VSI), virtual desktop (VDI) and databases. Their price competitiveness comes from de-duplication built into the architecture and advertised on their homepage as the “dedupe ticker”. At the time of writing, customers were averaging around 6:1 de-dupe rates and almost 13:1 with including thin provisioning. As Pure are targeting a wider market place, I think they will be more successful with their storage platform, however flash isn’t for everyone, so again, their market will be limited. In addition, at this stage their focus is in delivering consistently fast external storage, but we already know other vendors like Nutanix and VMware are looking to kill off the idea of a SAN with integrated storage and so diversification may be necessary.
The Architect’s View®
I hope that IPO goes well for both companies, however I do believe Violin will have more of an uphill struggle, simply because of their more targeted market. The flash market has matured massively in the last few years and diversification is definitely necessary; we see that evidence from Fusion-IO, plus the “big” storage vendors are already in the market and starting to push their own all-flash solutions. It will be intriguing to see how these two IPOs develop and which other companies feel the time is right to jump into public ownership.
- Violin Memory Files Registration Statement For Initial Public Offering (www.violin-memory.com)
- Violin Memory Launches 6264 Next Generation Memory Array (www.violin-memory.com)
- Violin Memory acquires GridIron Systems
- The Evolution of Solid State Arrays
- Solid State Arrays: Pure Storage Inc
- Pure Storage Closes $150m Funding Round (www.purestorage.com)
Copyright (c) 2007-2022 – Post #15FD – Brookend Ltd, first published on https://www.architecting.it/blog, do not reproduce without permission. Photo credit iStock.