I wanted to wait till the smoke cleared before sharing some thoughts regarding Microsoft opening the specifications on several of its APIs and communication protocols. Part of the released technology was the specification of RDP.
For VMware VDI, this is probably the most interesting of the released protocol specifications. The initial impact to customers may not be that obvious. However, over time, the benefits could be huge. RDP is the preferred protocol for accessing VMware VDI based desktops. Currently, this makes the most sense to leverage as there is no added cost in doing so which is typically the case with third party add on protocols.
Although getting the optimal performance in broad set of network conditions can be tricky, but it's quite adequate in the highest majority of desktop replacement use cases. RDP has always had virtual channels which allow a third party to add feature functionality to RDP. For example, with VDM 2 we developed a generic USB device redirector that allows you to connect most USB devices, not supported by RDP to a client access device and make them available for use with your VMware VDI virtual desktop.
In order for third parties to leverage virtual channels, you either needed a team of lawyers and a chunk of change or some really good engineers that could reverse engineer. For some time now there have been licensees of the RDP spec and others who reverse engineered it in effort to add value. This included people building feature add-ons as well as full RDP client implementations. Open Source efforts such as RDesktop and commercial efforts from companies such as WYSE, VMware, ThinAnywhere, Provision Networks, Sun Microsystems, Calista and others have all contributed to enhancing RDP.
Over three years ago I started researching the use of VMware virtualization as a way to better manage and deliver desktop computing and overcome traditional server based computing short falls. As we started meeting with customers and sharing the vision of VDI it was becoming clear Citrix had lost their seat at the table. Customers were concerned with the issues surrounding Citrix as a solution, such as complexity, application compatibility and loss of a personal computing experience.
VMware VDI was resonating with IT organizations and desktop management teams as an answer for more broad desktop deployments. I was a little surprised, as no one can take away Citrix's past contributions to server based computing. It's always been my opinion Citrix and Microsoft had become comfortable with their position and the direction they were leading customers and the market.
As VMware VDI emerged it was becoming clear this was what customers wanted. I had not seen such vast interest and adoption of a desktop solution in a long time. Despite Microsoft attempting to down play interest, VECD emerged. As Citrix lost ground trying to protect the Presentation Server stronghold, they finally broke down, did a 360 degree about face, and sank 500M on XenSource.
Since they have renamed their legacy products and are attempting to broaden their vision beyond just "Application Delivery". Here at VMware we are building things from the ground up with virtualization in mind, not trying to retro fit legacy solutions to fit in a virtual world. Since the emergence of VMware VDI, millions of dollars are going into the R&D of network delivered desktops and applications. The result is innovation that improves user experience, desktop management and choice for IT organizations.
What does this mean for Citrix? For years there have been the conspiracy theories that Microsoft will buy Citrix. Over the years I have seen less and less evidence or reason to believe this, but I guess it could happen. For some time now, I have said Citrix is getting squeezed from the top by VMware and squeezed from the bottom by Microsoft. The crown jewel for Citrix is ICA. Without that, there is not a lot there you can't get from other cheaper more simple alternatives.
With the acquisition of Calista by Microsoft, it's clear they intended to continue the enhancement of RDP. With the RDP spec now open, more and more engineers can come to the party and rapidly further enhance RDP improving its performance and features in a broad array of use cases. With the XenSource acquisition, Citrix has dipped a toe in the open source world. Now that Micosoft has openly published the RDP specification, will Citrix set ICA free? Do they have any other choice? What long term value does ICA have with so many big companies now able to enhance RDP and in fact, already doing so?
