Hey all my name is Gunnar Berger (@gunnarwb, http://www.gunnarberger.com) and I'm an End User Computing (EUC) junkie. I have lived and breathed in this space my entire career, although its only recently been called EUC. It all started during the early years of Citrix (WinFrame days) and has grown into what I foresee as the most important technology to date: Project AppBlast.
Let's talk about what is known about Project AppBlast, these bullets come from a mix of the VMworld press release and what my boy (whom I've never met) Scott Davis has shared in various videos you can find on the Internet
- AppBlast can deliver any application from any OS to any HTML5 browser
- No client is needed in the browser
- No additional licenses are needed for this (after all its just like webex... right?)
- The performance is good (but it looks like PCoIP outperforms it in this tech preview)
- It is not doing MMR (Scott Davis says as much in the video)
- One of lead architects of appblast is Jonathan Clark
- Jonathan was one of the guys behind a little project called... ThinApp... ya this guy rocks and is EXTREMELY sharp
Okay so the gist is, you have an application from any OS, AppBlast can send it via HTML5 to any HTML5 browser. Simple enough, so why do I think this is the most important thing since vMotion (or the iPhone).
1) Is it AppBlast or DesktopBlast
I keep saying this to anyone at VMware that will listen to me, but I get the same response. "AppBlast is about sending applications to any device." My feeling is that if you can send an application and it looks and reacts great, what's stopping you from sending the full desktop. Watch those videos with Scott Davis, one of the most difficult things to do well is video, that Video of Jonathan's trip to Antarctica looks pretty good to me. Another thing difficult to do well is USB, well it just so happens VMware already does USB well, its built into View, Fusion, Workstation and probably other products I don't use very much. So they have the tools its just a matter of putting them together.
2) MMR
MMR is great and its a PITA all at the same time. One of the reasons I'm such a fan of View and PCoIP is that its all host rendered which means very little configuration is needed. No need to update Adobe Flash every other week because Adobe fixed yet another bug, if it renders on the host it'll show up on the client. We see the same thing with AppBlast according to those videos. Scott is playing an outdated video format and the client can see it because it appears to be host rendered. I'd love to pick Jonathan Clark or Scott Davis's brain on how AppBlast works, for instance: What if that video was a more recent format, could it be MMR'd? After all, I'm already in a web browser, chances are I have a bunch of codecs in that browser. Could it run natively? Host rendering costs bandwidth but saves management, MMR saves bandwidth but costs management. Or to put it another way, one cost you OPEX but saves you in CAPEX and vise versa. I think AppBlast might be able to ride that line between the two, becuase its expected that you'd already have those codecs inside a browser.
3) Native HTML5 - No Client
The fact that you don't need to install a client to use AppBlast makes it possible for this technology to go viral. Just think of everything these days that has a browser in it, heck some fridges have browsers in it. What VMware is selling is the ability to send "applications" to anything. This is huge and would allow any client under the sun to start running AppBlast immediately. Also, thinking about repurposing old desktops, I'm positive you can find some linux USB boot stick that just gives you an HTML5 browser. Who out there is still waiting on the PCoIP Mac Client or Linux Client? No more waiting for clients, native HTML5 means Brian Madden can rest easy at night because his post about "its all about the clients" is answered with AppBlast: ANYTHING is a client!
Speaking of viral technologies. I'm pretty public about my love of Samsung and their embedded Zero Clients, they are trend setters over there. I was thinking that Samsung has something else called a SmartTV, it has a browser in it... If that can handle HTML5, guess what your TV just became a monitor. Samsung could start selling SmartTV inside of a standard computer monitor. Better yet, maybe DVI/VGA/HDMI cables become a thing of the past. A lot of TVs these days are equipped with WiFi. If I can send a desktop over Ethernet, and my TV has a HTML5 browser and Wifi all of a sudden the DVI/HDMI cable becomes obsolete. Who doesn't want less cables!
4) Terminal Services:
I have no idea how this technology works so I'm just going to throw this out there... what if it could work in conjunction with Terminal Services. I mean based on how Scott talks about the technology it just forwards whatever you tell it to, could this be true for TS as well? This would be amazing because View doesn't really compete with XenApp because XA runs on TS and as such has a much lower TCO (if you don't include the cost to hire someone like me to fix it when it breaks
) . If AB just became a plugin for TS, and VMW bundled it with their USB redirection technology, well, Citrix stock might drop a few digits.
Conclusion:
I think AppBlast is about apps for right now, but I see a lot of potential in this technology. I think if its priced right we could see the complete changing of how monitors and desktops interact. Even you home PC or Mac could just run AppBlast and you wouldn't have to bother plugging it into a monitor, on top of that any TV in your house could become your monitor without any special cables or having to run wires inside the wall so your wife doesn't yell at you for making the living room a mess... The point is, look out for this technology, it could change everything.
References:
http://www.livestream.com/vmwarecommunitytv/video?clipId=pla_4b440ad4-e336-4419-9cfa-b7476a3bbc85
http://blogs.vmware.com/euc/2011/08/vmworld-2011-introducing-projects-appblast-and-octopus.html
