I was having a conversation last week with some folks about the evolution of VDI and how the New Desktop will evolve to run anywhere, including locally and offline. Someone asked a question which was essentially "If your June 2010 VDI vision becomes real, then is terminal server or server-based computing (SBC) dead?" It's a great question, and the short answer is "No. Both TS and server-based computing will be important in the future, even in a world full of VDI and client hypervisors."
The future of Terminal Server in a VDI world
If you believe that VDI will truly become as easy to manage as TS (with single disk image management, etc.), and if you believe the user density will of server-hosted VDI will get pretty close the user density of TS (measured hardware in cost per user), then I think it's pretty safe to assume that just about all hosted desktops will move to VDI instead of Terminal Services. After all, if cost and management was not an issue, then why wouldn't you use VDI for all desktops? In this case, a TS desktop might even be more expensive because that would be a completely different kind of technology you had to support. You would have the choice of all users running a client OS (just some local and some hosted), or local users running a client OS and hosted users running a server OS. Which would you rather choose?
This means people wouldn't use Terminal Services for remote hosted full desktops. It does not mean Terminal Services is dead, as people would still use it as the underlying OS feature to serve remote seamless SBC applications. (In other words, all desktops, even remote ones, would be some form of VDI, but many remote apps might still be on Terminal Services.)
The future of server-based computing in a VDI world
I've talked a lot about the fact that VDI and Terminal Server are both server-based computing. (link to video about this) So when I talk about the future of "SBC", I'm talking about remote computing / SBC in a single user (VDI) and multi-user (TS) context. Today's SBC has four main advantages:
- Management (all users are in the datacenter)
- Access (use Windows apps from non-Windows devices and from outside the firewall without installing or streaming them)
- Performance (client-server apps work great, since the remote application client is in the datacetner with a fast connection to the application server)
- Security ("eyes only" security, with no real data ever leaving the datacenter)
These are the same four advantages of SBC that Citrix has been touting for over ten years.
In the future, the evolution of VDI into the New Desktop will remove the "management" advantage of SBC (in that we will get management for all computing scenarios, even offline / local clients). But the other three--access, performance, and security--will still be advantages of SBC, even in the future.
So even in a world where all desktops are virtualized and managed from a master image and streamed to local hypervisors... even in a world where applications are packaged and virtualized and streamed on-demand to local users... there will still be scenarios where you need to use a fat app over a slow connection, or regulatory compliance issues where no datasets can leave the firewall, and for those cases, good old server-based computing (be it a multi-user or single user back end) will be your friend.
Even though I can't wait for the evolution of VDI and the New Desktop, I say "Long live TS! Long live SBC!"