A week later than some people predicted, the news has broken. NetApp has bought SolidFire for $870M. This continues the trend for established storage companies to acquire start-ups with great flash products rather than building their own. SolidFire initially targeted the service provider market with its scale-out all-flash array. In the last year or so, SolidFire has taken aim at enterprise data centers as well. In both of these markets, NetApp is well established as a vendor of disk-based storage arrays. However, it has been losing market share due fundamentally to a lack of a modern solid-state product. The FlashRay product that NetApp developed has failed, leaving NetApp no option but to acquire.
NetApp doesn’t have the best track record for acquisitions. This one appears to be its largest and is definitely critical to future success. I believe that a key success principle will be to keep the SolidFire product engineering separate from the existing FAS lines. Customers deploying all-flash arrays want products that are designed to get the most out of flash, without any legacy disk-based thinking. Trying to shoehorn SolidFire into a FAS architecture is likely to be painful—as painful as clustered ONTAP coming from the Spinnaker acquisition.
On a personal note, I have a lot of friends at SolidFire. A larger company buying a start-up can easily fall apart because the cultures are different. I hope NetApp treats the SolidFire team as a success and keeps its start-up ethics and approaches. I hope it also tries to keep the SolidFire team together rather than integrating everything into existing NetApp units. Further, I hope the company is revitalized by the new ability to address the solid-state opportunity.
NetApp’s purchase of SolidFire is likely to be the last acquisition in the IT infrastructure industry in 2015, but we can expect some more activity in 2016. I have to think that Cisco is looking at the market. Its problems with the Whiptail purchase will make it a bit gun-shy. If it wants to compete against the Dell/EMC machine, then it needs storage. I think Cisco needs to make a couple of storage acquisitions to show the market it is serious. A hybrid array, an all-flash array, and a scale-out hard disk array would make a portfolio. Of course, Cisco could just buy NetApp and get a full storage portfolio, which would then include a great all-flash array. Buying NetApp has the advantage of a huge number of loyal customers, with storage teams that want NetApp products. NetApp has a market capitalization of around $8 billion. It would be a big acquisition, but nowhere near as large as Dell’s purchase of EMC.
-Disclosure NetApp Employee-
Hey Alaistair,
As the APAC Flash guy for NetApp, part of my job now is to explain both internally and externally where each piece of the flash portfolio fits into the evolving datacenter puzzle. Although I’m limited about what I can say about Solidfire and integration plans until the deal actually closes, what I can say after the “aha !” moment about where soldieries’ strengths complements the data fabric capabilities we’ve already developed, a lot of the real excitement comes from bringing in a lot of people who live and breath the idea of building service oriented infrastructure. There are already a group of good people within NetApp who can do this, but not nearly enough of them. It’s not just about technical and soft-skills, it’s about mindset.
Solidfire isn’t just another Flash Array, in some respects from my point of view the fact that it’s made out of Flash is almost beside the point. What impresses me is that they built something that is genuinely disruptive because it should change the way people consume technology. Of course it’s made out of all flash because it simplifies the design and makes QOS much much easier to implement. Disks in many respects as amazing as they are from a technical perspective, are complete nightmares when you try to do anything resembling predictable performance.
All the other All Flash Arrays in the market – including Pure, XtremeIO, 3PAR, All Flash FAS and E-Series aren’t really _that_ disruptive. They’re mostly excellent products, but they all have a pretty simple value proposition – i.e – They let you “do the same dumb thing you’re doing now just cheaper and faster”.
Maybe for some vendors and customers there is some process change because its the first time they’ve been exposed to working versions of “Instantaneous Copies/Snapshots/Clones” and “Storage Efficiency/Thin Provisioning/Compression/Dedup”. For NetApp though, that’s all stuff we’ve been using as differentiators for about a decade, or two decades if you include snapshots . NetApp already has a culture focussed on turning those features into something that helps to transform traditional datacenters into something more modern. A lot of that work involves helping people bridge the gap between being “custodians of infrastructure” into “managers of a service”, but that’s a slow process for many involving managerial and culture changes that most pay lip service to, but face significant hurdles before they can truly walk the walk. For those people on that journey (which is probably about 85% + of the market today) ONTAP and the NetApp skill set that goes with it are an excellent solution, and helping people on that journey to hybrid cloud is and will remain a core part of the ONTAP design centre.
From my point of view Solidfire’s design centre is completely different, it is for the people who have already made the transition to full ITaaS, or perhaps avoided the transition by starting fresh. Furthermore the Openstack expertise within Solidfire is just damn impressive, NetApp was already doing pretty well in that area and had made a lot of investments in both personnel and technology, but in combination with Solidfire it will mean there will be two major options for.. openstack storage deployments .. CEPH and NetApp .. after that its just also-rans.
Oddly enough I actually disagree that NetApp has a poor history of acquisitions, it’s true that EMC outbid us for both VMware and DataDomain, but thats a luxury you have when you have deeper pockets. Judging on the same basis the Dell purchase of EMC is doomed to failure. StorageGrid, Engenio, Altavault (nee SteelStor) are all developing nicely, though most their true potential still lies in the future of IT.
Also, it’s true FlashRay is dead, but not quite in the way most people think . FlashRay wasn’t killed by SolidFire, it was killed by All Flash FAS. The work we did in ONTAP to absorb the technology created by the MARS team meant that NetApp was close (very very close) to having two deduplicating scale out All Flash Array products that had only minor differences. Sorting out the positioning and go to market for that situation was, to say the least, problematic. Having said that the technology created for FlashRay lives on, its mission of helping NetApp create a world class, high performance All Flash Array for the evolving datacenter has been fulfilled. There’s more to come, but that technology won’t be packaged up as yet another all flash array (YAAFA) in a crowded market.
Merry Xmas, Happy Holidays, Seasons Greetings and my good wishes to you and all your kith and kin.
Hopefully we can catch up for a beer sometime in the new year
John Martin
Hi John, I hope you are enjoying some good weather for the holidays.
As always your perspective from inside NetApp is interesting. NetApp seems to be one of the companies that has passionate staff who believe in their products. I agree that the transformation to IT as a service is going to be long and slow. For vendors this means a long but declining tail of selling and supporting legacy products. I think it is clear that designing products for new service delivery requires a redesign from first principles, not an update to a ten or twenty-year-old product. This is true for hyperconverged as well as storage. These new designs happen in startups, which are then acquired. I think the relatively early acquisition of SolidFire gives NetApp a good chance of succeeding. So long as it is clear to NetApp that SolidFire is the future, not just a feature.