Cathy Spence in Intel IT has a great blog on the specific use case that she ran with System Manufacturing that starts in my mind to ask the question around streaming both OS & applications. Here's her Blog @ http://communities.intel.com/openport/blogs/client/2008/02/14/could-streaming-apps-run-faster-than-traditionally-installed
The question that I have is that application streaming technology has been around for a long time, Windows Terminal Server environments are a NORM.. so why does OS & application streaming really raise the bar on the new computing? or is this just a matter of time when the HW & SW will become a trusted partner beyond the standards today to a new more compelling future?
what do you think? what makes streaming compelling or....../?
ECMF Blog
Previous Next
paulghahn
None of these solutions (server-side application virtualization, client-side application virtualization and OS Streaming) are meant to be exclusive. Each exists to better enable an IT organization to match the delivery paradigm for a particular application with a set of end-user characteristics. Got a traditional client/server application (say SAP ERP)that needs to be accessed by administrative users or remote users? Then look at the benefits of using server-side application virtulization (Microsoft Terminal Services or Citrix XenApp (nee Citrix Presentation Server)). Have a windows application that a mobile employee needs to use while on an airplane? Use client-side application virtualization (application streaming). Have a high-security or classroom environment where you need to re-set the OS image repetitively? Use OS streaming.
The point is that there isn't any one single right approach. But but rather there are now available a range of tools available to the IT organization for application delivery.
adbatra
I agree with paulghaan that IT now has several options for delivering applications. The challenge or job for the IT is to understand the implications of each delivery method and implement the most suitable compute model or delivery option for each application. This choice can be granular at an application level rather than a user level. All applications for a user may not be delivered by the same delivery method and a user does not necessarily use teh same compute model all the time. For example a user may use a high security application through terminal services only when connected to the network but use a different application offline on the plane. Two different applications, two different delivery methods (two different compute models) but same one user and same one client-device i.e. note book.
Just have to give my 2 cents here. I spent a fair amount of time setting up thin client environments with Terminal Server, just to realize I shifted the burden from the clients desktop to profile and system management on the server, where I actually had a higher rate of service tickets than for rich clients.
For me application streaming holds more promise. It seems much more manageable, or at the very least it's more flexible. I guess you could stream the whole experience if you wanted to get to something like the thin client model.
The advantage here is this is more like the Cable TV service model. You sign-up for the package of stuff you will use on a regular basis on your client PC, but when you need a little something extra you download that on-demand (if this avoids setting up another profile on Terminal Server that's worth it alone). Being able to run an alternate OS, or get premium apps when you need them seems pretty cool and very flexible for the workforce (getting AutoCad, PS or 3DS when you need them.. just too awesome) while taking the burden on IT service desk.
Also I think this is where applications are going as we see more apps move to the cloud and off the desktop. Could be a way for IT to get an architecture underway that prepares their environment and culture for this model.
So it seems pretty cool to me.