VMware VSAN: When good enough more than passes muster

Last week I attended SFD9 and the last session of the week was with VMware about VSAN 6.2. This is not a review of the product or an analysis about single features; there are plenty of them already. Instead I’ll talk about what a product like VSAN 6.2 means for the entire SDS/HCI market.

6.2

The new version was launched a few weeks ago and it’s an evolution of the product originally launched less than two years ago.

VSAN follows vSphere version numbering, which means this is actually version 2.2 and not 6.2. But this is not really important.

The first version (5.5) was really 1.0 – almost disappointing. Version 6.0 added basic features missing in the first release. 6.1, released six months ago, seemed like a different product altogether because of the number of improvements and features added. The latest version, v6.2, is another step forward in terms of overall efficiency and features; something you should pay attention to.

If, at the beginning, it could have been considered as a catch-up play in the VSA/HCI space, it’s pretty clear now that VMware has caught up with all of the other suppliers, by which I mean all the startups in this market. Even with its limits and constraints, VMware VSAN looks solid and consistent now. Though it might not be perfect for all use cases, you can’t take it into consideration if you want to build a vSphere-based HCI.

Youtube Video

It’s an undeniable success

There is no doubt that VSAN is a success. With its 3000+ installations (paying customers) it’s the most installed VSA for vSphere in the market after HP's StorVirtual VSA. And I have no doubts that the numbers will continue to rise thanks to the maturity of the product. If at the beginning it was just good enough, now it is a very good option and it is part of an end-to-end solution.

In fact, the real question is “why not?” Well, you could be sceptical about some technical details and still-to-be-improved features or the absence of multi-platform support… but still, if this product was originally designed with the goal to hide all the complexity of storage from the poor “jack of all trades” who runs your virtual infrastructure (aka the work-buried VMware sysadmin), then it’s doing a fine job.

And this is bad news for many startups in this space. I don’t know for sure, but it seems it’s become tough to beat VSAN from the TCO standpoint now and TCA could easily become a viable measure – especially when commercial bundles, discounts, ELAs and so on come into play.

VSAN, EVO:RAIL, VxRAIL and TCO

VSAN has been a success from day one, but EVO:RAIL is a totally different story – a complete failure.

And VSAN doesn’t directly compete with HCI vendors (like Simplivity or Nutanix, for example). Products like EVO:RAIL, and now VCE VxRAIL, do. I really hope that VxRAIL or the evolution of EVO:RAIL in general, if any, will be more competitive with these vendors. And yes, in this case also from the TCA standpoint!

I have no particular preference here, but I still see Nutanix and Simplivity as more sophisticated solutions, especially if you want to work with a vendor that doesn’t own the full stack and you want to have options to run/move workloads across different hypervisors with the same management layer.

But again, if you are a VMware-only shop – VxRAIL? Why not? In time, for competitors it will become harder to find a differentiator. I think it is too soon to look for TCO comparisons, VSAN 6.2 and VxRail are very new on the market after all, but I’d like to see something soon.

Closing the circle

VMware is doing a good job with VSAN and the sales speak clearly. Part of those sales could be fanboys, others could come from ELAs (sometimes it’s just easier because the procurement process)… But that’s not all; there is much more and the latest VSAN release is there to confirm that VMware believes in this product.

While I still see a clear differentiator in some of the competing solutions out there, for many of them competing becomes difficult. From this point of view, for example, Springpath did the right thing in coming to an OEM agreement with CISCO, but life for the smaller players will get consistently tougher. ®