Impressions on the App-V User Group day 2011

23.11.2011

App-V

So, in the last week there was a first (European) App-V User Group day, that was held in Microsoft NL’s offices at Schiphol / Amsterdam. Like almost a 100 others, I was there to attend the event and even though I did not have my own presentation, I got my share of publicity as well.

First of all, I have to say that the event was very well organised (though there could have been cold drinks – like water – in the morning before the sessions in addition to the usual coffee), and thanks for that goes to Login Consultants who acted as a host to the event. But in addition to Login, an equally important factor was of course the presentations themselves for which I should say thanks to all the individual presenters!

The morning kicked off with the presentation from Ruben Spruijt, who presented on the subject of Project VRC (Virtual Reality Check). While the VRC itself is not particularly App-V specific in any manner, but rather a series of benchmarks and whitepapers (divided into phases) which aim for the understanding of the various moving things that goes into building a well-functioning VDI system, the phase IV of the VRC concerned application virtualization in VDI as therefore was very much on topic for the day. Besides, I can hardly speak for myself when saying that all the information presented by Ruben in his session was of general interest, not least because reading through number of whitepapers – as good as they might be – is always bigger task than be presented the same in good presentation. Sorry Ruben for that, but sometimes we are lazy ;-)

Second presentation after the Ruben’s was coming from two people from the Microsoft itself, Madelinde Walraven and Sebastian Gernert. Madelinde works for MS in the Netherlands and Sebastian in the Germany, and both are in the support side so they apparently get all sorts of interesting customer issues related to App-V into their table. The topic of their joint presentation was about troubleshooting techniques and commonly occuring issues. While I cannot say that I personally learned all that much from the session (maybe besides being reminded – once again – about looking more closely into Process Monitor and how to best utilize it), I’m sure that many in the crowd did.

One thing that kind of stick to my mind from that presentation is that the art of the App-V logfiles does not seem to be common knowledge anymore. That’s at least how I interpreted the reaction to the part where those logfiles were discussed; in the “old days” Softricity SoftGrid training guide used to list those logs extensively, and people in the courses were taught about them. I guess not so much anymore..? Of course things like a NTA log has always been mostly-undocumented feature of App-V and not present in the official training materials. As are the HTTP -streaming related logs btw, which was perhaps the only genuinely new thing to me.

What was, however, another rather interesting part in Madelinde’s and Sebastian’s presentation was the treatise of DSC and how horribly people abuse the feature. While I’m sure that the word “AppCloud” now brings smile to everybody’s face who were in there, it does not seize to amaze me how inventive people tend to be when they build their App-V environments. If you have over 10000 TCP socket connections to a server, man, you seriously have a problem which is not the fact that App-V Management Server runs out of it’s dynamic pool of ports. On the other hand, this highlighted the fact that the whole RTSP streaming in App-V is architected (just) a little bit wrong, since it requires that many ports per package be constantly open.

After having a coffee break, sessions continued with Ment van der Plas‘ presentation about App-V support in the upcoming Configuration Manager 2012 (which is still in beta). Like I think Ment mentioned during the presentation, some of the topics for this integration were heavily debated during this year’s MVP Summit back at the Redmond and sadly the integration is still not 100% perfect from App-V’s perspective. Ment did a very good job of highlighting the good and the bad, as well as clearly spelling out what the integration looks like in practise. Like with previous presentation, there wasn’t too many new things to me as I had seen the SCCM team’s presentation at the summit already, but since I have not really played with SCCM2012 in hands-on manner it was never the less a good reminder of how all the pieces fit together. And most importantly, how much better the new management interface is as compared to that MMC thingy of the current version..

I still think, btw, that the whole SCCM is bit over-engineered solution in architecture-wise (look at things like RES Automation Manager (nee Wisdom) for how to make a lean infrastructure) for most of the cases, but it is going into better direction. It’s also too bad that the SCCM folks clearly aren’t App-V experts and thus do not build all of the things into their product that would benefit from App-V package deployment perspective (like DSC linking that Ment mentioned); the mindset is too much on the traditional software distribution still, even with the all the fancy evaluation(?) rules and whatnot.

Fourth session in the day was one which was presented by Nicke Källén and was on the subject of heavy-duty sequencing. Or should I say rather that it was on the topic of busting the commonly accepted “no, can’t do that with virtual app” truths. Of course, some/most of the limitations are actual limitations of the technology – such as the shell extensions – but Nicke seems to be rather hell-bent on finding way to do those things anyways in conjunction of virtual applications ;-) This, naturally, means playing all sorts of tricks with the locally deployed components, which may or may not suite your particular situation. Some of these things has already been documented on the internet (like how to create those shell extensions) and some of the things can be worked around with ready-made tools. And some you just have to know. Interesting discussion overall, I’d say.

During Nicke’s presentation, I was summoned on stage to say a few words on the topic of SFT file 4GB limitation. Like I remarked back then, I felt like an expert witness bought to trial on this subject, but I was really honored to have the opportunity to share my views on the matter I happen to know very well. Now, during the brief discussion we had while being over there, I theorized with Nicke that 4GB could be perhaps worked around with a clever DSC -based trick, so hopefully we have some sort of continuation for that train of thought at some point. But before we are there, I still would disagree a bit with Nicke’s assertion that you could actually have fully working, single, over the 4GB SFT package since there are pointers that will overflow. Small FB1 or not.

After having a lunch break and some socializing with the other attendees (seems that there were lot of people either already using our commercial App-V tools, or thinking into using them [yes, you should!] :-)), the afternoon started with the presentation about Server App-V -product. The presentation was given by Jurjen van Leeuwen, who was the only one of the MVPs present I had not met before. Jurjen, btw, also moderates Appvirtguru.com forums so I want to thanks him for that as well. Since Server App-V is integrally tied to the System Center Virtual Machine Manager (SCVMM) -product, there was also a brief demonstation of that as well as the local Powershell -based usage of the Server App-V cmdlets. I have played with some early version of Server App-V in the past, but from the looks of it the version shipping with current beta of SCVMM has been improved a little bit UI-wise so maybe I should take that into more close examination once again. And yes, from Jurjen’s demo I again got confirmation on my old impression that SCVMM is heavy, UI -wise.

While the Server App-V as a product is obviously related to the desktop version, it’s still quite a different beast for couple of reasons: For one, it has decided to forgo the isolation of the applications, from what I understand. What this primarily means, and this is purely my opinion, is that we are not really talking about virtualization that much anymore here. The files are laid out in the filesystem and it is really a alternative packaging/deployment mechanism for an server application. What’s more though is the fact that Server App-V packages contain configuration items, which allows you to prepare the environment (like adding new users or groups to local account database) and/or configure the application (like changing database connection string for the app). These changes, for most part, are actually changing and configuring the local machine and not virtualizing those items per se.

Little bit like the whole MDOP tie-in in desktop App-V, App-V on servers is limited to only customers using SCVMM as I already mentioned. Personally I think that Microsoft should consider releasing the technology to be used on all Windows servers, not just ones managed by SCVMM. Even if we are not talking about strong isolation of virtual applications like with desktop version, I could clearly see benefits of having the technology available for general consumption. For what it’s worth, from marketing and sales perspective I can understand the reasons for having this connection like with oh so many Microsoft technologies these days, but still it would make sense to have Server App-V as standalone technology as well, even if it’s against a little price.

The day continued from the Server App-V stuff back to a desktop version, this time concentrating on a Dynamic Suite Composition, or DSC, part of the App-V. This presentation was done by Falko Gräfe, who was among the first App-V MVPs back in the day and has then been re-awarded later on after being away for couple of years. It was very interesting presentation, although a little bit “schematic heavy”, but Falko did a thorough job of covering not only how to use DSC but also potential conflicting states and pitfalls when suiting together multiple packages. Which you should definitely avoid if there’s even a remote chance of having conflicting registry or file entries between those packages! I’m sure that many people in the audience learned a lot of new things about DSC and what they should and should not to do, as well as having a good overview of 3rd party tools available that makes creation and visualization of DSC links more easy. All of them were familiar tools to me (in one way or another), but especially add3’s one seems very handy one. And speaking of add3, at this point I should probably apologise to Danny that I never managed to send those testing materials for the DSC tool which was causing failure on my machine ;-)

The last “real” session of the day was done by Rodney Medina, a guy with even longer App-V history than I have. While Rodney is not apparently not doing much of App-V stuff himself these days, he of course has extensive knowledge and experience from practical issues of many years. The topic of the Rodney’s session was not purely App-V -only, but rather touched App-V on appropriate places while discussing the general topic of User State Virtualization – or profile virtualization, depending on which way you want to look at it. The whole concept of USV is interesting one, and I think we haven’t really solved the problem in full even with the latest tools. As somebody (maybe Tim Mangan, who has a clear interest towards this “data virtualization” puzzle) said it at some point, maybe only Microsoft can solve this thing completely as they control the keys to the kingdom, namely Windows profiles and what applications should/can do on Windows.

Rondey and his company, Immidio, try to solve user state portability from their perspective by enabling roaming of application and user settings independent of the physical user profile, but as Rodney laid out in the presentation there’s also things besides the settings (like user’s documents etc.) that need to be handled. App-V itself does some of the isolation of settings through the PKG files (and no, the post series wasn’t about all you and your mother should know about PKG files but rather about VFS, which relates to PKG of course ;-)), but as pointed out in the presentation, there’s overhead associated with storing things inside PKG. All in all, I’d say that like Ruben’s presentation first in the morning, Rodney’s one was also interesting specifically because it played on a general level and not being strickly App-V only in content. Sometimes (most of the times?) we need a broader context against which to judge App-V specific nuances.

The offical part of the day ended with open “ask the experts” – or should I say, “experts ask from you” – type of session, where all the presenters (plus me) were brought to the stage together with specific “claims” on screen that acted as discussion starters. These were things like “App-V and security” and “Packaging is hard”, commonly stated myths that everybody tend to have their own opinions on. The idea was that myths would be either busted or confirmed, which I think played out well in the end.

Still after the official agenda, there was open-ended drinks for all event but it seemed that most of the people were in hurry flying out from the Amsterdam (hardly a trivial task, due to extensive fog situation). So there wasn’t maybe as much conversations going on as there would have been if more people would have stayed longer.

Overall, a good event and I do hope that it will be repeated next year with maybe bigger venue (as promised) and more attendees!

,

About Kalle Saunamäki

As one of the first four Microsoft App-V MVP's, Kalle has been doing application virtualization since 2003 and virtualization in general from 2000, and is a recognized in-depth technological expert in Microsoft application virtualization community.

View all posts by Kalle Saunamäki

Subscribe

Subscribe to our RSS feed and social profiles to receive updates.

Trackbacks/Pingbacks

  1. What everyone and their mother should know about VFS in App-V, pt. 3 – A special leak-through condition for fully virtualized folders | Gridmetric Blog - 20.12.2011

    […] Another, secondary, benefit of using fully virtualized folders is to confine all data that the application may store on that folder during runtime; all possible changes to a fully virtualized folder will be 100% cached in App-V Client PKG and won’t be passed through to the local system. But as has been discussed, sometimes this can actually be a problem since you might want to let some type of files go to the local system, like database files that could potentially grow into too big of a size for App-V PKG to handle (around 1GB for the whole PKG file, according to Sebastian Gernert when this matter was touched in a bit different context during last App-V User Gro…) […]

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: