When VMware got really serious about View at VMworld 2009 with the PCoIP announcements, VMware took several interesting positions:
- That Desktop Virtualization via VDI (or hosted virtual desktops – HVD) was the obvious next generation way to do end user computing and that HVD was going to replace traditional fat client desktop computing with VMware View leading the charge.
- That HVD’s were different and superior to terminal services based solutions like Microsoft Remote Data Services (formerly Microsoft Terminal Services) and Citrix XenApp (formerly Citrix Presentation Server). This superiority was supposedly due to the fact that every user got their own virtual OS (as opposed to sharing a server OS), and because you did not have to deal with multi-user applications issues and cross-applications compatibility issues.
Since 2009 several things have become clear as a result of how various products and approaches have fared in the marketplace:
- The use case for HVD’s is not in any dramatic way broader or different than the use case for terminal services. In other words, there really is not any significant or broad objective that you can accomplish with an HVD approach that you cannot accomplish with terminal services and vice-versa.
- VMware offers a point solution when it comes to the desktop computing space and Citrix offers a broad range of solutions most of which are well integrated into an applications delivery suite. For VMware View is the product and the strategy. For Citrix XenDesktop is just one feature in a suite that includes terminal services (XenApp), application streaming, application virtualization, and a type-1 client side hypervisor (XenClient).
- Whereas PCoIP was a clear improvement over RDP, it was not a clear improvement over ICA/HDX especially in WAN scenarios. The bandwidth requirements of PCoIP lead many organizations to conclude that it was only useful for LAN attached users which basically ended the idea of it being the remote protocol that would allow centralization of all of the desktops.
- While VMware has done well with View, Citrix has done better with XenDesktop. Citrix has done an excellent job of leveraging its installed base and its partner channel to build a growing XenDesktop business on top of its XenApp business. VMware has suffered from organization turmoil in the leadership of its Desktop Virtualization unit, and has suffered from the embarrassment of not begin able to integrate the profile management assets acquired from RTO Software in a timely manner.
All of the above coalesce into the following situation. VMware has five businesses; data center virtualization (vSphere), cloud computing platforms (vCloud), applications platforms (vFabric), management software (vCenter) and desktop virtualization (View) These businesses are discussed in detail in this post about VMware’s New Stack. Of these 5 businesses, View is probably #5 in terms of priority to VMware. This is primarily because the other four businesses have strong cross-synergies and because VMware has figured out that it is not going to rebuild the Citrix product line in order to compete with Citrix. Citrix has one main business which is the delivery of managed desktop computing and applications via a suite of products of which HVD is just one feature or option.
So what about that partnership? Back in February 2010 we asked the question of whether a Partnership Between Citrix and VMware was Feasible. That post analyzed the question of whether some formal deal between the two companies was possible that was based upon Citrix supporting PCoIP and perhaps the two clients getting merged into one. While this level of formal and technical integration has not occurred something more important has.
What has occurred in the meantime is much more organic and pragmatic. Citrix has basically given up on the sever virtualization business and basically just makes XenServer into an installation option for people who want virtualization but who do not already have either vSphere or Hyper-V installed. VMware has given up its grandiose ambitions taking over the entire desktop computing market and wiping out Citrix in the process. The marketplace has realized that HVD’s are probably not going to get very far beyond the 15% of enterprise desktop penetration that Citrix and Microsoft achieved with terminal services. And here is the important point. As a result of all of these pragmatic realizations, Citrix has been implementing over 50% of its XenDesktop installations on a VMware vSphere back end, and VMware has realized that this is good thing.
In other words, while the marketing departments at the respective companies have not gotten together and issued a press release proclaiming a new day of marketing, sales and technical cooperation, the new day is already here. Both companies have realized that customers want the market leading application delivery suite (Citrix) to work on the market leading data center virtualization platform (vSphere). The partners who implement both solutions (VMware and Citrix have many common VAR and Integrator partners) are interested in having both companies succeed, and are more than willing to integrate the two solutions. VMware wins because an important new class of workloads (XenServer and XenDesktop) gets implemented on its back end driving further sales of vSphere. Citrix wins because they do not have to fight vSphere with XenServer to get XenApp and XenDesktop sold. So there you have it, VMware and Citrix are partners – they have just not announced it.