VMware Virtual Volumes Launch A New Storage Era

The arrival of VVOLs with vSphere 6 will allow us to provide data services at the VM-level and put an end to impractical LUN management.

Howard Marks

February 12, 2015

4 Min Read
Network Computing logo

VMware’s Groundhog Day announcement of vSphere 6 means that I’m finally released from the non-disclosure agreement that was part its Fight Club-style public beta program. Just as the first rule of Fight Club was to never talk about Fight Club, the beta program was open to the public but participants were sworn to secrecy.  

From my perspective as NetworkComputing’s steely eyed storage guy, vSphere 6 with Virtual Volumes (VVOLs) represents a sea change in the way we view and manage storage. In the very near future -- VMware says this quarter -- we’ll be able to stop managing LUNs and start providing data services at the virtual machine level.

This will overcome the difficult relationship virtual platforms like vSphere have with most storage systems. For various reasons, including SCSI/Fibre Channel/iSCSI protocol limitations, most VMware installations have a relatively small number of storage LUNs provisioned to the vSphere cluster with multiple VMs in the datastore on each.

If you take a storage system snapshot in that environment, it snaps all the VMs in that datastore, making the snapshots huge, and since you can’t quiesce all the VMs at the same time, that snapshot will be crash, not application, consistent. VVOLs give the storage array the data context to provide those services at the VM-level, or actually the VMDK-level.

This functionality becomes especially important with quality of service. Today, if I put my vSphere infrastructure on an array that supports QoS like a SolidFire or NexGen, I can say this VM needs the platinum level of service and put it in a platinum datastore on a platinum LUN. That will protect that VM from noisy neighbors at other service levels or even other platinum VMs in other datastores, but because the storage array isn’t aware of what data belongs to which VM, it can’t enforce QoS against other platinum VMs in the same datastore.

Particularly interesting is VVOLs' SDN-like separation of the data path and command/control paths between the ESXi host and storage array.  Data goes over Fibre Channel, NFS or iSCSI just as in the pre-VVOLs storage dark ages. VASA 2.0 provides the command/control channel that passes storage system capabilities up from the array to vCenter and commands to provision, snapshot and otherwise manage the VVOLS from vCenter to the array.

As I reviewed VMware's announcement, I also was glad to see that the storage industry is climbing on the bandwagon with support. A  dozen vendors that were VVOLs development partners, including HP, NetApp and Dell, along with,  young bloods like Nimble, Tintri and Atlantis Computing promise VVOL support in the first half of 2015. Another group, including VMware parent EMC, have promised VVOL support at some later date.

I take the fact that EMC is in the later camp as evidence that poppa-EMC doesn’t micromanage little lord VMware as closely as some would profess. If making poppa happy was No. 1 on VMware management’s to-do list, VVOLs would still be delayed.

With VVOLs and VSAN (which gets significant updates in vSphere 6), VMware is clearly stating -- as I have in the past -- that we have to shift the unit of storage management, and services, to the workload. Microsoft, Citrix, and Red Hat should consider themselves challenged to bring per-VM storage management into their hypervisors.

For more detail about how VVOLs work, read my blog post "The VVOLs Cometh," check out the vBrownbag session I recorded or listen to the episode of the Greybeards on Storage podcast where my co-host Ray Luccasi and I chat with Satyam Vaghani, who did much of the early work on VVOLS when he was at VMware.

Last year, I had grown impatient waiting for VVOLs, and compared the technology to the never-appearing title character of Samuel Beckett's Waiting for Godot. So to close, let me coin the line that Estragon never gets to say: “Godot enfin arrivé.” Godot, or at least a VVOL, finally arrives.

Attend the live workshop, Making Cloud Storage Work for Your Organization, given by storage guru Howard Marks at Interop Las Vegas this spring. Don't miss out! Register now for Interop, April 27 to May 1, and receive $200 off.

Disclosure: HP, Dell, Microsoft, Atlantis Computing, SolidFire, NetApp and EMC are or have been clients of DeepStorage, LLC.   Tintri provided a T540 storage system for use in DeepStorage Labs.

About the Author(s)

Howard Marks

Network Computing Blogger

Howard Marks</strong>&nbsp;is founder and chief scientist at Deepstorage LLC, a storage consultancy and independent test lab based in Santa Fe, N.M. and concentrating on storage and data center networking. In more than 25 years of consulting, Marks has designed and implemented storage systems, networks, management systems and Internet strategies at organizations including American Express, J.P. Morgan, Borden Foods, U.S. Tobacco, BBDO Worldwide, Foxwoods Resort Casino and the State University of New York at Purchase. The testing at DeepStorage Labs is informed by that real world experience.</p><p>He has been a frequent contributor to <em>Network Computing</em>&nbsp;and&nbsp;<em>InformationWeek</em>&nbsp;since 1999 and a speaker at industry conferences including Comnet, PC Expo, Interop and Microsoft's TechEd since 1990. He is the author of&nbsp;<em>Networking Windows</em>&nbsp;and co-author of&nbsp;<em>Windows NT Unleashed</em>&nbsp;(Sams).</p><p>He is co-host, with Ray Lucchesi of the monthly Greybeards on Storage podcast where the voices of experience discuss the latest issues in the storage world with industry leaders.&nbsp; You can find the podcast at: http://www.deepstorage.net/NEW/GBoS

SUBSCRIBE TO OUR NEWSLETTER
Stay informed! Sign up to get expert advice and insight delivered direct to your inbox
More Insights