This discussion has been locked.
You can no longer post new replies to this discussion. If you have a question you can start a new discussion

What Is The Future of Application Delivery?

Every year seems to be met with an optimistic glow of it being the “year of the desktop”. By that, I mean that a large shift will push virtual desktop infrastructure (VDI) into the lime light. Virtualizing a desktop is pretty easy, and lots of products out in the field can wrap a desktop OS inside of a hypervisor and provide brokerage. This is often the sticky spot, too, as the goal shouldn't be to simply “make your desktops virtual” but to completely shift the delivery method to embrace an improved desktop model. I've seen some deployments fail because a traditional model of patching desktops, doing antivirus scans, and pushing out software were laid on top of a virtual infrastructure and caused crippling failure.

So let’s assume you have provided the hardware infrastructure to deliver the performance needed for a virtual desktop. Again, not terribly difficult to do with today’s modern technology. The question then becomes – how do you deliver the applications? Recall that the entire point of a desktop is to provide a platform in which the user can connect with their software. Is the answer thin apps, streaming apps, or continue with thick installed apps? I tend to think this will be a combination of the above.

emoticons_plus.png Have you embarked on the journey towards providing virtual desktops, or even thin / streaming applications, to your clients or users? Care to share your experience and opinion?

emoticons_plus.png What challenge have you met that had to be conquered, compromised on, or caused your project to halt or shift in a new direction?

Reply to this post and earn 1 entry to win an iPod Nano!

  • WE have been virtualizing apps in our environment for a couple of years now.  Delivering virtually Office, RSA administrator and several others with a roaming desktop.  Wherever you log in, you apps are waiting.  We are now looking at moving our lower end users ( meaning those with a base image) to a virtual desktop.

    The one issue we have run into is older applications do not seem to like being virtual.  We actually even have a couple of applications that are so finicky, that not only are the desktops still XP but the servers that back end the can not be moved to the cloud.

  • Call me Old School but I still do prefer to try to Jam as many applications as I can into a 'Golden Image' and then sort out the experience with the user via Group Policies and other 'On the Fly' customizations programmatically as the user logs in.  Obviously Application conflicts and temperamental apps will force a layered approach but there 's nothing like the speed and integration that a locally installed application gives to the user.

    CARLO.

  • From my own personal experiences with setting up a few smaller VDI environments, applications can be a tricky thing some times. Most of the ones I worked on, we were using VMwares VDI solution with Orchestrator and the client had a few different VM templates that had specific software installed for that type of user or group so that a persons AD group would determine what image they received when logging in from their thin client. I would also think a citrix type application could work, then you would just need a VM with the citrix client on it and the person logging in will have access to all of their stuff. This would allow you to maintain your apps in a central location and have a way to control and update your desktops easier (by updating the master image and "rebuilding" all of the VMs built from ssaid image).

  • We mainly just keep the 'golden image' updated with apps but there are times when we do thinapp.  My preference would be to get to the point where more than 50% of our apps virtual.  We are slow going with this only because there are (in my opinion) many apps that just don't virtualize well.  In addition, we were told by a consultant that heavily relying on ThinApp would decrease the login time of our View desktops.  I'm not sure if this is really true or not, but if there is any chance that ThinApps would slow down the user's login time (even if it's only once a month when we refresh the VMs) we would opt out.  Login time is a critical component for us and one of the main reasons that we have switched to desktop virtualization here at the hospital and our surrounding clinics.

  • I think the point is what is the FUTURE of APP delivery...As with all technology there's exists a learning curve to understanding the technology and also a curve to actually delivering. Many times theses curves can happen simultaneously where learning and delivery can be accomplished at the same (or near same) rate. I feel VDI is not one of these cases....The return to processing at the core was, in my opinion, a natural evolution back to where computing began, but the recent hurdles have been what can be delivered to the end users efficiently (green screens are not the desired app). Currently I think the technology is mature enough to provide the backend support for VDI and scales very well, but there's still strides to be made in adapting the software suites to function properly in this environment. With the popularity of Cloud computing, virtual/home offices and hosted networks I see apps to be forced into the virtual environment and IS/IT staff learning to support them in this way.

    My feeling is that within 2-4 years this will be a moot topic as the world of VDI and streamlining apps delivered in that architecture will be exploding. Maybe the Topic question will be "What did we do before VDI". Until then it will definitely be a messy mixed bag of methods!

  • We had some early conversations about moving towards a VDI solution; however, we ultimately have decided not to go in that direction for now.  In our environment we don't have a many users to there wouldn't necessarily be a huge benefit.

    As an outsider looking in, it doesn't seem like there has been a big movement from the application vendors to support VDI environments.

  • We actually sell VDI where I work (network engineer so don't mess with it to much) and in all honesty it seems to have really had a positive impact on all of the clients that I have interacted with. We are a full Citrix environment in regards to that all the way down to the hypervisor. I personally really enjoy the experience of being able to access all of my content from a cloud environment when I am not at the data centers, it just makes sense to me.

    As for challenges it would probably plugin integration's at times and getting homegrown applications published to a VDI or XenApp environment.

  • Yeah, legacy applications can be a real sticky situation. If there's enough commonality between users, there's a potential to bake them into the gold image by commonality (or the "shotgun" approach to distribute to all). This can be made complex with licensing for some apps.

  • Your'e so old school! (You told me to call you that, so I did). emoticons_wink.png

    In a linked clone type deployment, having an army of apps in the golden image can be done - or with a storage array that deduplicates when using full "thick" images. I'm definitely a big fan of GPOs as that's what I used to use "back in the day" to push apps. Today, though, there's so many more ways to do that - streaming, thin, or even remote app (think: XenApp) that it becomes quite interesting in how you design the app delivery.

  • Application packaging is a big ticket item from a level of effort perspective. Especially when you have an army of apps (I typically see 400+ apps in most deployments) to try and package. Most of the time a thin application won't slow down the login time - ThinApp, for example, can stream what is necessary to kick off the app in a "just in time" fashion, and only delivers the rest as needed. But obviously our mileage may vary depending on the app. Some apps, like Office Suite, are ones that I prefer to just bake into the image more often than not for this reason.