Tom Fenton recently published an article on Virtualization Review detailing the current state of VMware’s new Virtual Volume (VVol) storage architecture. In the article he polled a few vendors to find out what they are seeing as far as customer adoption of VVols. A few vendors responded including myself, both HDS & Dell did not have an accurate way to track adoption and where mainly relying on customer feedback. They are mainly seeing customers testing it out right now and using it in Dev/Test environments. HDS stated one of the limiters to VVol adoption is customers still on vSphere 5.5 and Dell stated customers are still trying to understand it better before diving in.
At HPE, we can track actual usage of VVol adoption via our array phone home capability which provides us with some usages stats on the array. In the article based on my feedback Tom wrote that we had seen at least 600 3PAR arrays with the VVols VASA Provider enabled within the array. More recent numbers puts that at around 720 arrays, but its important to note that this just means they have the potential to use VVols, not that they have VMs running on VVol storage. More detailed stats show that about 50 customers have created VMs on VVol storage. So this is pretty much inline with what other vendors are seeing which is pretty light adoption of VVols right now.
VVols has been available as part of vSphere 6 for almost a year now (March 2015), so why aren’t more people using it? There are probably a lot of reasons for this including:
- Customers haven’t migrated to vSphere 6
- Array firmware doesn’t support VVol
- Lack of replication capabilities in VASA 2.0
- Lack of knowledge/understanding of VVols
- Limited scalability and feature support in some implementations
- It’s essentially a 1.0 architecture
In my previous post on when customers would start adopting VVols I went into a lot more detail on the barriers/challenges to VVol adoption. I expect usage to pickup within the next year or so based on a number of factors:
- VASA 3.0 with replication support in the next vSphere release
- More arrays support for VVols
- Increased scalability and more feature support
- More mature implementation from VMware and array vendors
- Better understanding of VVols and how to implement it
Until then I expect to see steadily increased usage of VVols, like any new technology or feature, adoption is almost always slow at first as customers are often cautious about jumping right in to something new. The same growing pains were apparent with VSAN as well when it was released as a 1.0 new storage architecture. If your array supports VVols I encourage you to definitely try it out and learn all you can about it as VVols is the future and at some point I expect VMFS to be phased out just like ESX was. If you are looking for resources to learn more about VVols be sure and check out my huge ever-growing VVols link collection and also my VMworld 2015 STO5888 session that VMware has made publicly available.