While most of our industry is funded by venture capital, it should be no surprise that much of the talk around the cocktail table revolves around mergers and acquisitions. The question I get asked most frequently is, "Which storage vendor is Cisco going to buy?" While the answer to that question might be interesting, I think the more interesting question is: “How will the vendor horse race affect the products I can buy?"
Over the 15 years or so that I’ve followed the storage industry, there's been a common pattern of product development. A group of vendors would pioneer a new market, say, iSCSI disk arrays for SMBs or deduplicating backup to disk appliances. Then, once the startups reached a $50 million to $100 million annual run rate, or a $2 billion valuation, the established vendors -- primarily EMC, Dell, and HP -- would snap them up.
Established vendors would acquire, or develop, solutions for the newly defined product category, and the few startups left when the music stopped -- like Xiotech and Nexsan in the wave of SMB disk arrays -- had a tougher row to hoe.
However, today we have a more vibrant set of innovative storage vendors than at any time I can remember. From all-flash and hybrid arrays to hyper-converged solutions, new companies are challenging the established crowd with better technology and lower prices. At the same time, the usual buyers don’t have the gaping holes in their product lines that led to the acquisition of companies like Data Domain, LeftHand, EqualLogic, and Isilon.
Sure, you may argue that Nimble or Tegile has a more elegant design that makes more efficient use of flash than a VNX, but the net effect is that EMC has to configure the VNX with a bit more flash to get comparable performance. Each of the established vendors has a “good enough” all-flash and hybrid product. They’re dabbling in hyper-convergence with OEM deals like Dell/Nutanix, Cisco/SimpliVity, and, of course, support for VMware’s EVO:RAIL program.
The good news is that the VC community is willing to give today’s startups plenty of runway. Companies like Pure Storage and Nutanix have hit the $2 billion valuation mark in their latest funding rounds. In the past, the VCs would have been yelling, “Take the money!” in the founder’s ears or behind his back.
Then there’s the Cisco question that started off this whole discussion. Just as Cisco had to expand into the server market once it sucked all the growth out of its core networking market, eventually UCS will rise to its natural market share, and Cisco will need to expand into the storage space. Some of my fellow pundits have wondered why Cisco hasn’t bought a storage vendor already.
To me, it’s pretty clear that Cisco’s UCS server division has gotten a huge boost from the industry trend towards having vendors, rather than IT architects at the user organization or a VAR, choose the detailed configuration of systems.
Since Cisco was the one brand-name server vendor that didn’t have a storage division, whenever a storage vendor wanted to create a reference design or a converged computing solution, the servers in that solution would be Cisco’s. From VCE to Nimble, the storage vendors use UCS in their solutions because a Dell, IBM, or HP-based product would allow those vendors' storage salesmen into the deal.
I think the overall value of that market boost is still big enough that Cisco is better off waiting 18 to 24 months before attacking the storage market directly. At that time, I can see it buying one or two software-defined storage solutions and shifting them to UCS platforms to cover the middle of the storage market. But before taking another bite of the storage pie, Cisco will have to learn the lessons of its Whiptail acquisition, which has not gone all that well, with Cisco suspending Invicta sales.
The recent approval of Lenovo’s offer to buy IBM’s x86 server division could throw a wild card into the whole market. If the folks over there, like Michel Dell and the management of HP, want to be Thomas Watson’s IBM -- the soup-to-nuts supplier of IT goods and services -- they’ll need some storage. I expect they’ll take a year or two to absorb the IBM server folks. But 2016 could be a big year for storage M&A.