I’m surprised that The Register, with its humorous yet poignant headlines, didn’t run an article titled something along the lines of:
NetApp to VMware: “EVO is nice for branch offices and stuff, but leave the heavy lifting to us”
Apparently, the whole NetApp EVO:Rail announcement took VMware by surprise. Duncan Epping, Chief Technologist at the VMware CTO office commented in his blog, “Although I have been part of the EVO:RAIL team, it is not something I would have seen coming.”
From a datacenter disruption standpoint, the EVO:Rail partnership is important because it indicates yet another of the “big 7” have now announced their own hyper-converged solutions; only IBM and Hitachi remain without an offering (not counting standard EVO:Rail for Hitachi). But I have my doubts about how serious NetApp actually is:
- NetApp has publicly stated, “FlexPod works more in the enterprise data center and large offices, while EVO: RAIL is more for department and branch office deployment outside the core data center.” I can just imagine that the VMware folks are grinding their teeth about that quote.
- Adding a Filer, or any SAN/NAS storage, kills the EVO:Rail scale-out story – one of the most powerful attributes of a hyper-converged architecture. In other words, once customers fill up the Filer, they’ll need to purchase another Filer.
- EVO:Rail isn’t cheap. And even if an organization has a VMware ELA, it must still purchase the EVO:Rail licensing on an OEM basis from the manufacturer. When it is time to upgrade the hardware, the licensing must be purchased again. Adding NetApp will, of course, make the solution still more expensive.
- There is confusion about what the offering really is. No one even knows which servers will be used (best guess: Lenovo or Fujitsu). One thing is almost for certain, it will be complex. NetApp and VMware are probably banking on VVOLS with policy management to help administer the environment, but VVOLS itself is not yet proven.
- Since NetApp cannot compete with a truly hyper-converged solution, it is trying to move the EVO:Rail architecture back toward the FlexPod/Vblock architectures by adding capabilities such as data deduplication, compression, cloning, replication, etc. But it will be difficult to message the NetApp EVO offering in respect to FlexPod. Support will likely be challenging (is it a VMware EVO or NetApp issue?), flexibility will be limited, and resiliency constrained by the RAID and other archaic options of an array-based solution.
NetApp appears to have rushed this announcement to market – it didn’t want to be left out of the hyper-converged revolution. I suspect that while NetApp may use its EVO:Rail offering to open doors, that its reps will still primarily be pushing FlexPod.
Time, of course, will tell whether I’m right or totally off-base. In the interim, I would be very interested in hearing from readers, especially from channel partners and potential customers, about your take on the NetApp EVO:Rail announcement.