Flashing upstarts tempt tech titans' throbbing disks
If you can't beat 'em, buy 'em instead
Blocks and Files Should mainstream disk array vendors buy the flash startups, which are going to take their lunch money, now or later?
Spend time with Kaminario, Nimble Storage, Pure Storage, SolidFire and Violin Memory, drink the Kool-Aid, hear them talk the talk and see them walk the walk. You'll think to yourself: "This is solid stuff, pure storage gold. There is no way Dell, EMC, HDS, HP, IBM and NetApp and their mature disk-based arrays can compete against these flashy startups and their clever, clever software. They are are going to eat the major vendors' lunch."
The basic reason is that disk array software is mature, multi-layered, complex, tuned to disk and multi-functional, although it generates millions of dollars of revenue a year. There is no way a flash array can be slid underneath this software and still perform as well as an enterprise-class flash-based design that's been built from the ground up.
The first take-out will spark a feeding frenzy
OK, then the majors could decide to build their own, but they then face formidable problems. There is a restricted pool of talent in the software and hardware disciplines involved, and the cream of it is already working for the startups with stock option-plated handcuffs. It's hard to attract effectively locked-in talent.
The startups are working closely with flash vendors and are busy tuning their kit. Say EMC or IBM builds its own all-flash array product and introduces it in 2015. The startups are already in the market, taking business from the majors and looking to at least $1bn-a-year revenue streams. Customers have already decided to buy their kit, and, if the majors have compatibility between their new flash arrays and their existing product lines, it will take longer to build the product and may affect its performance.
Let's take it as read that the flash array and flash-enhanced array products are real and will succeed. Then the majors are in an unenviable position: build or buy. They will surely do what they have consistently done with significant startup technology advances over the past decade or so and buy.
It's a virtual foregone conclusion. The only question that remains is when? They can either buy now at pre-IPO and before a rival product ships, spending between $200m and $300m, or they wait to see which little biz is successful while losing business to the startups and then buy them for much, much more, like $750 million to a billion or more, perhaps as much as $2 billion.
The fast and smart will buy now. The cautious and clever will buy in 2014 to 2016. My betting is on the majors taking the startups out around 2015. Going a bit further I'd assess EMC as moving first and HDS last with Dell, HP, IBM and IBM buying after EMC but before HDS.
It's a high-stakes flash array casino. There are a restricted number of tables and more punters than tables. The first take-out will spark a feeding frenzy. Who are going to be the winners? ®
Sponsored: RAID: End of an era?