On March 11 on the VMware end-user computing blog, Sumit Dhawan casually announced VMware’s next big thing: Project Enzo. A former Citrix exec, Sumit Dhawan is now VMware’s senior vice president and general manager, desktop products, End-User Computing.
Choosing the day before your chief competitor’s annual conference to mark the public debut of your next big thing is becoming standard operating procedure for Citrix and VMware. Last year, Calvin Hsu, Citrix product marketing VP for all things XenApp/XenDesktop-related, swore that it was purely a coincidence when Citrix announced XenApp and XenDesktop 7.6 immediately ahead of VMworld, but he wasn’t fooling anyone. Now it is VMware’s turn, and as preemptive announcements go, it looks like VMware has got the better of Citrix this time. Where everybody knew that new versions of XenApp and XenDesktop were in the works, the Project Enzo announcement has come as a surprise to many.
Sumit would have you believe that Enzo is “a next generation hybrid cloud architecture designed to transform virtual workspaces.” That sounds impressive, but I can’t say I’m 100% convinced by his description. To me, Enzo looks to be both a lot more and a lot less than that. A lot less because there’s precious little Workplace in Enzo, a lot more because tucked away in the announcement is the introduction of a new technology that VMware hinted at last year—one that, if it works, will significantly change the way we do VDI.
Let’s get this Workspace business out of the way to start with. Enzo, according to Dhawan’s blog post, “is designed to combine the economic benefits of cloud-based VMware virtual desktops and application technology, with the simplicity of hyper-converged infrastructure.” Sorry, Sumit, but desktop and apps don’t make a workspace, not anymore. Citrix CEO Mark Templeton took time out of the opening keynote at Synergy last week to clarify just what it takes to make a workspace:
What is a workspace? Most of our colleagues in the industry are promoting it as desktops as a service. Basically, it’s a desktop. That’s what a workspace is. We think that’s the old definition of a workspace.
Is a workspace a desktop? Yes.
Is a workspace a Windows app? Yes.
Is a workspace a mobile app? Yes.
Is a workspace a collaborative space? Yes.
Is a workspace a web app? Yes.
A workspace is whatever you need, whatever kinds of digital tools you need to accomplish your tasks independently of the kind of device, in fact optimized around the device.
–Mark Templeton, President and CEO of Citrix Systems, Inc.
I side firmly with Templeton here. Right now, Enzo is a VDI play, so whatever it is, Enzo is not a workspace.
But what it is is much more interesting, and far more deserving of Templeton’s attention.
Project Enzo (any comparisons with similarly named exotic Italian automobiles are purely coincidental) is both VMware’s answer to Citrix Workspace Cloud, and some really cool tech that is going to change how VDI is done across the board. Citrix introduced Workspace Cloud at the Synergy conference last year at the Anaheim Convention Center, when it went by the name of Citrix Workspace Services (reviewed here and also covered here in an interview with Citrix’s Juan Rivera).
Citrix is billing Workspace Cloud as:
“…a new control plane that merges the worlds of on-premises and cloud, allowing IT managers to create secure, mobile workspaces that include desktops, applications and data from whatever infrastructure source best meets their specific needs.”
Contrast that with:
“The new VMware Enzo Cloud-Control Plane, a cloud-based management layer hosted on VMware vCloud Air, is designed to offer IT administrators a ‘single pane of glass’ to unify the administration of on-premises or off-premises virtual workspaces.”
And you’d be hard-pressed to tell the difference between the two.
Workspace Cloud is Citrix’s next big thing. It should have given Citrix breathing room to differentiate itself from VMware and chart a path for the future into less-contested waters. But it’s not yet ready for prime time, and now it looks very much as though VMware isn’t going to give Citrix the opportunity to create a new, uncontested niche for itself. It’s difficult to attempt any meaningful comparison between the two products right now. Workspace Cloud is obviously significantly more advanced than Enzo in its current state of development, but it’s clear that there is going to be significant overlap between the two. While Citrix has a twelve-month advantage over VMware at this point, it is by no means certain that this lead will result in significant sales. Enterprise customers are naturally cautious and are likely to be slow to embrace this new architecture, giving VMware plenty of time to catch up with Citrix before sales start to take off. Even then, it is by no means clear that Citrix and VMware’s enthusiasm for cloud control planes as alternatives to old-school admin consoles will extend to enterprise customers. And this time around, both Citrix and VMware are entering this market behind start-ups like IndependenceIT, which is continuing to strengthen its already strong cloud desktop orchestration credentials, and Virtual Bridges, which recently sold off its VERDE DaaS platform to concentrate its efforts on orchestration services. I’ll take a closer look into each of these competing orchestration services in the next few weeks, but for the moment I’ll just say that all this talk of cloud control planes and orchestration only serves to distract from the real meat of Project Enzo.
The other half of Dhawan’s blog post explains how the Cloud-Control Plane links back to something called “Smart Node,” which will reside on the future (hopefully less woefully underpowered) EVO:RAIL and EVO:RACK hardware. VMware Smart Node will act as the glue that binds the appliance to the cloud and takes care of all of the orchestration actions. If Smart Node only turns up on EVO:RXXX systems, it will have achieved one major step forward for VDI just by being there. If nothing else, Smart Node will cement the role of the hyperconverged infrastructure appliance as VMware’s preferred means of delivering virtual desktops. I’ve been saying this long enough; perhaps now this means that someone is listening. Unless you are a service provider doing cloud-scale VDI, hyperconverged appliances are the only way to go.
But even Smart Node is still just a distraction. The stuff that really matters today is VMware’s Instant Clone, aka VM forking, aka Project Fargo, which enables you to clone (and personalize with a unique GUID, MAC address, etc.) a running virtual desktop (or any VM, for that matter) in no more than a couple of seconds. Instant Clone is copy-on-write on steroids. Boot one golden image on each appliance, clone it 100 times over, personalize it with AppVolumes and User Environment Manager, and there you are—instant desktops. Not only that, but Instant Clone blows a hole in all the current IOPS challenges that have dogged early VDI implementations. As a quick reminder, VDI needs IOPS because booting Windows 7 needs IOPS. However, if you clone a desktop after it has been booted, the demand for IOPS evaporates, and so does the need for that expensive all-flash array. Yes, VDI can still benefit from fast storage, but VMware Instant Clone repositions flash storage from being a necessity for delivering acceptable performance to what it should have been all along, a means of delivering virtual desktops that are faster than physical desktops.
A qualified public beta of Project Enzo will be available later in the summer of 2015. Register at Ready for Project Enzo to Change Your World? if you are interested.