So the Mystic/Marvin cat is out of the bag and VMware has announced EVO:Rail as their entry in the Hyperconverged infrastructure space. There are several great blog posts about this announcement, and you can tell that VMware was very selective in which influencers they targeted and how much access they were given prior to launch. I would say go read this post from @ChrisWahl on the WahlNetwork as it has a very concise high level overview. I would follow that up with Part 1, Part 2, and Part 3 of a series that Julian Wood has put together. And of course I would be remiss if I didn’t link to Duncan’s posts as well given his new position as Chief Technologist at VMware.
Now at first blush, my initial impressions were in line with my thoughts around what could VMware do given the amount of time that VSAN was available, and how long would it take them to qualify the hardware. Bottom line: this is a quickly crafted HCL compliant platform with an simple to use deployment and management engine. The VSAN ready nodes, were an easy telegraph that VMware was going to be moving in this direction, and it gave them some good feedback in how to craft a solution (as well as how to pick an HBA Controller). As I see it, the special sauce here is indeed the management and deployment portal, everything else is just VMware. The time to market required to build something like this isn’t significant, and VMware has been able to execute quickly on bringing a product to market. Kudos to them for entering the Hyperconverged ecosystem. It’s certainly a validation of what Nutanix and SimpliVity have been working on for the last 5 years.
In the big picture this move certainly means that Hyperconvergence is here to stay, and in my view you will see over the next few years a large shift towards a more converged compute/storage datacenter. It will eventually displace the traditional multi-vendor Converged Infrastructure stack (VCE/Flexpod) though not right away (still looking at the 3 year time horizon there). As the major OEM’s decide on their strategy of how to enter this market it will be interesting to see new alliances that are going to be formed, as well as a number of acquisitions that I think will be happening in the next 12 months.
Now its off to the races for the major vendors to put their special twist and value add into either existing products that are being warped to meet the new space, or build a new solution from the ground up. Nether of those efforts are trivial, but I would comment that anyone building their own stack from ground zero, even if they are just tweaking common open source software, will have a bigger challenge. Nutanix has gone this route, but as early entrants into the space, and with 24 months of sales and customer interactions under their belt, they are probably the last group who will do this full stack from the ground up approach. SimpliVity has also done a lot of development of their own, and in conjunction with their custom hardware acceleration for data virtualization, have an approach that is more HCL driven as it takes time to validate their software/hardware approach with a commodity X86 stack, i.e. see the current SimpliVity/Cisco relationship announced yesterday. Both companies have ample experience to draw from and a significant head start over new entrants. Still, neither can compete with VMware when it comes to install base, marketing muscle, and lets face it, available funding to push this solution. Those are the face cards they can play today as they ramp up.
In the short term, the advantage goes to Nutanix and SimpliVity, primarily because they have been selling into (and creating) this market for the last two years. Both Nutanix and SimpliVity have had to deal with the challenges and the learning curve associated with creating a market, and there in lies the benefit for both groups. The experience gained from fighting in those trenches cannot be purchased, bought from marketing campaigns, or created in the lab, it has to be earned on the battlefield. EVO:Rail at first blush seems awfully restrictive, and as someone who spent 18 months in this space, I think there are some assumptions that were made by VMware that will preclude this product from getting broad traction or support in some of the verticals they have identified. Of course pricing is TBD, so that will be interesting to see as well given what I know about the costs of the other players in this space. We always see the product first, then the price a few weeks later. I still see a strong need for a platform that provides HA, 1-2TB of storage, and the ability to run 5-20 VM’s at a price point of around 20-30k as a huge sweet spot, especially for large retailers and groups that have 100+ remote sites/offices to address, but I digress, and that will be a post for another day.
For now, lets Welcome VMware to the Hyperconvergence club. I wish them much luck, I think they may find this space to be much tougher than they expected.