AppDetective - All Comments

re: The "beginning of the end" for XenApp based desktops, prayers for layers, and can VMware "judo flip" Citrix Flexcast?

appdetective — Tue, 30 Aug 2011 02:43:51 GMT

Ok I'll break my own rule and send out a quick update during my break. I'm flattered by the responses so here is one final update and then I must really sign off and go into occasional view mode.

@Ron, valid point simpler is not always better, but application compatibility is a huge concern for me. I already know App-V is not a broad solution for me, only good for a subset of apps and I question everyday why bother as it does not go far enough. I don't expect Citrix to do anything much with management of layers any time soon and that's where Unidesk remains ahead. No reason not to implement Unidesk in a single layer mode and build from there is it makes sense and the apps works. I hope Citrix still says it is ok for XenDesktop customers to use Unidesk and they are ok with it. I doubt it though, and perhaps you need to build upon their layers format and just extend and all will be well. I would not be holding my breath for VMware to do anything as all early signs from VM World today show they will do nothing apart from keep telling a post PC story with that Horizon which is little more that a powerpoint, plus seems I'd rather write cloud web portal myself with API access. As for VDI vs. TS, if VMware does not keep pushing VDI that will kill the market for Citrix and put them in direct conflict with Microsoft who will only push RDS. Not good for Unidesk, but how funny would that be. VMware Judo flip Citrix back into the cross hairs of Microsoft and they will become an obedient B I T C H once again or Quest will start to get more MS field love once again kiss their ass.

@Dougdooley, thanks, I focus on real world implementation don't care much for the vendor side as this gives me less access to a broad set of technologies from many places where I can pick and choose what I want/need to build real business solutions. Not knocking your suggestions, but just not my thing plus I would just end up telling too many people to F off they are stupid since they know nothing about the realities of how to implement which is huge gap on the vendor side IMO and a real shame. I'm tired of empty suit execs who know nothing about how to get things done trying to tell how they know best. There are of course exceptions on the vendor side and you seem very pragmatic but that is rare. Hopefully Cisco can do some M&A in our area, hopefully I gave you a few ideas to consider.

@SillyRabbit, thanks I can please some, each to their own. I say what I feel mostly without filter unless I blog and just try to share my experiences. I bet Citrix is selling XenApp under the XenDesktop umbrella and nothing more. If VDI was that good, I'd expect VMware to have a lot more examples of successful implementations in broad markets even with the protocol gap. Don't understand you layers and control comment. With Unidesk you have a admin model, with RingCube you have little so a weak crutch.

@Issy sure centralized management is fine, but I don't buy a totally new approach in a world where there is already a lot of sunk cost in existing tools and process and limited time to get to Windows 7 is going to get much traction. We have been down the image management path before, Ghost, Altiris etc, and we all know from experience it get's you so far, a lot more granularity is needed in a dynamic user community. I prefer your physical and virtual approach to Unidesk, but for them that is just current state. However both of you face the same problem IMO, it's a total bet on a management layer. Bare to try and a hard game to win given time limits.

@Navot, in the mobile space, I believe native experience wins because that's what the developer will use. I think Gizmox has technology to reduce the need for XenApp published apps more than I believe it will make any serious dent in the "local" mobile app world.

@harrylabana Sure fair points, I don't disagree that both vendors and customers need commercial incentive and the big guys for the most part just build crap that can't be implemented and then take years to do anything and even then its mostly crap. I am ok paying as you put it for value. The problem is not enough value is created vs. the marketing and it's too slow or low quality AKA AppSense 8.0. I also think the big vendors adding some features to put fire under the feet of the eco system is a good thing as it raises the bar, so a good balance. So agreed, technology that enables the business has value that I will pay for. However that value needs to be maintained and not just the usual vendor defense arguments of switching costs are high so they won't. This is why I like the SaaS model. Crap vendors can be just switched off. I'd like to consume more technology this way as the vendor switching cost defense will be removed.

re: The "beginning of the end" for XenApp based desktops, prayers for layers, and can VMware "judo flip" Citrix Flexcast?

harrylabana — Mon, 29 Aug 2011 15:01:40 GMT

AppD I was forwarded your post my a number of people while I was on vacation so I guess you struck a cord with a few people. Certainly you have a good appreciation of User Virtualization and that is good to see. Overall a decent article although I'll refrain from commenting too much on some of your theories about my former employer and simply say that don't assume there is a conspiracy to do any one thing. Lot's of things happen at the same time, which may lead to a conclusion at a point in time. Over time these things tend to work themselves out. I'll leave it that.

More importantly I wanted to comment on an assumption that is prevalent in your post. The assumption that I'll assert you are making is that consolidation is always a good thing. The theory would go Citrix/VMware buy a layers or user virtualization vendor and make them features that customers pay nothing or very little for and things all of a sudden just start to get implemented and everybody is happy. That can certainly be true for some features. However there are several factors to consider. What is the lifespan "as is" of a feature, integration effort and how much additional investment will it take to make it good enough/sustainable for most people? I'll bet as we have seen time and time again, there is limited investment unless there is direct revenue to justify increases and accelerated investment. What that results in is slow progress for many sophisticated customers who want more features and faster innovation. You complain about Citrix features that you would like to see all the time faster. Imagine if as people have speculated over the years, Microsoft bought Citrix. Do you think Microsoft would invest as much in Citrix as Citrix do or would it just become a better RDS and stay relatively the same for several years? This is why there is an independent eco system that innovates faster to fill white space in the industry and also why customers vote with their wallets to validate the eco system to fill their needs. It's also why many customers hate paying software maintenance, amplified with large vendors as the continuing value get's harder to justify for a broad customer base and you end up with good enough, safe bets. Innovation is risky and expensive, good enough makes enough money for the quarter….

You also talk a lot about lock-in. Lock in to a vertical stack. From my own customer experience I totally get that and the desire to build component architectures that allow you to switch out architectural components as better or more cost effective options become available. It's a key differentiator for XenDesktop that allows you to work across many hypervisors although vendors love lock in as it means lot's of revenue and less reason to accelerate risky innovation or expensive M&A. I'd argue user virtualization vendors don't force you to pick a hypervisor, a desktop broker or a particular client architecture so you have choice, but of course we'd love you to lock-in to our user virtualization stack :-)

It is also clear from my experience that customers don't buy user virtualization with desktop virtualization budgets. As you point out correctly, many large, complex and very sophisticated customers use AppSense at scale today. They buy our wares in addition to their desktop virtualization investment and increasingly as part of their Window 7 migration budget that also applies to the majority desktop use case, physical desktops . They push us to the feature limit, that requires lot's of investment on our end and if you ask them they will tell you that they want us to go even faster. That's not going to happen as a free feature, or the ability to support such organizations which is very expensive as you scale. Customers want capabilities that add value and the confidence that their investment will continue . They pay for value even when they desire to have it cheaper and faster. That tension is healthy, value for the customer and incentive for the vendor. Value + incentive = progress. Free feature usually equals low value or point in time value with limited forward investment driven by low incentive resulting in incrementalism.

re: The "beginning of the end" for XenApp based desktops, prayers for layers, and can VMware "judo flip" Citrix Flexcast?

Navot — Sat, 27 Aug 2011 07:23:50 GMT

Great reading. Adding layers is adding complexities; I believe it is like saying the globe is round. We @ Gizmox look at the discussion here from a web perspective. We were able to flatten endless web layer architecture into one layer in term of programming languages, methodologies, practices etc and we claim that it is doable here too. Having doing that, we still maintain the classic MVC model which makes life so much simpler. I would say that the only path to simplicity is the path of layers diluting. And this were we should be heading. We do it by Incorporating Gizmox Automatic apps Transposing, employing one layer-based web as a network platform in order to extend virtualized application reach and enabling the consumption of such application via browser and form-factor aware agent. The beauty of all, is that while thinking out of the box, we remain absolutely standard. The innovation is under the hood, surface is as standard as it could ever be. I believe that being standard, simple, single layer could change a thing or two in the virtualization space. My 2 cents.

re: The "beginning of the end" for XenApp based desktops, prayers for layers, and can VMware "judo flip" Citrix Flexcast?

Issy — Fri, 26 Aug 2011 18:47:23 GMT

Nice and well thought article... Layering is not a panacea, and can easily get out of control if mis-used or over-implemented. For us at Wanova, the focus has been on decoupling the managed image from user-installed apps so IT can manage their enc-users workspaces without forcing them to lock-down...

As for your comment on Wanova being a "pure management play" -- Isn't it all about trying to offer a better mousetrap for desktop/workspace management ?? (as opposed to being "hypervisor-bound" ?)

The advantages of desktop virtualization are in centralizing images for single-image management, DR and business continutiy, fast break-fix, Win7 migration... What if this is achieved w/o requiring a hypervisor (although absoloutely supporting it if its there), supporting uniformly both physical and virtual, central or remote endpoints ?

re: The "beginning of the end" for XenApp based desktops, prayers for layers, and can VMware "judo flip" Citrix Flexcast?

SillyRabbit — Fri, 26 Aug 2011 18:23:36 GMT

@AppDetective - Thank you for a great posting. I for one enjoy your rants and cussing since it put color into the conversation and gets people engaged!

@All

Why does Citrix have to fix this? Who is their target customer?

Answer - the Enterprise customer. Only the folks who can afford to host desktops in their datacenters.

Call it VDI, HDV, mashed potatoes. Even resellers (Citrix Partners) are having a tough time selling XenDesktop without the license conversion option.

Why is layering important?

Answer – It gives perceived control back to various organizational stakeholders rather than solely on IT. It is a crutch to facilitate organizational buy-in.

It’s sometimes silly how products have to be sold.

re: The "beginning of the end" for XenApp based desktops, prayers for layers, and can VMware "judo flip" Citrix Flexcast?

Ron Oglesby — Fri, 26 Aug 2011 14:53:17 GMT

DougDooley

I dont know about the glove box. I have plenty of "rants and the F-bombs"... just dropped one yesterday when I as at the end of my rope for a moment. What I had to develop was a filter. My first few attempts were pretty bad because I might have stopped some of the language but body language told the story. But a simple filter vs a lock box is a wonderful thing. You can always open up the flood gates when needed!

I cuss like a sailor when I am out fishing. Maybe its just my proximity to water and being on a boat at the point. :-)

But you're right. there is some insight here and AppD's thoughts (when put together like this) can be useful even if you dont agree with everything.

re: The "beginning of the end" for XenApp based desktops, prayers for layers, and can VMware "judo flip" Citrix Flexcast?

dougdooley — Fri, 26 Aug 2011 14:34:38 GMT

App D-

Agree with @Ron Oglesby - "well thought out article" and an admirable amount of restraint considering the length. ;-)

Some really good "food for thought" - thank you.

You might think this is a crazy idea but the way your mind is expressed here is... you could help shape markets/industries by working in a corporate development role at a large tech company doing M&A. The other route is on the VC side of the deal but the barriers into that world are pretty substantial.

Obviously, the rants and the F-bombs would have to get "locked away in the glove box" more often that you might like but you've proven in this piece you're capable of doing just that so.. something to think about as a use of your analytically talent. Since I don't know who you really are OR where you work - maybe that's your regular job doing M&A. If so, then great but something tells me that's not the case.

Well, enjoy the time off...

Doug

re: The "beginning of the end" for XenApp based desktops, prayers for layers, and can VMware "judo flip" Citrix Flexcast?

Ron Oglesby — Fri, 26 Aug 2011 13:18:59 GMT

Oh... and on your thought of Desktops vs XenApp...

TS/RDS will be around for a while, But I couldnt agree more with what you are seeing. The push/direction seems to be directing everything to to the new model in the long term.

re: The "beginning of the end" for XenApp based desktops, prayers for layers, and can VMware "judo flip" Citrix Flexcast?

Ron Oglesby — Fri, 26 Aug 2011 13:16:53 GMT

Good reading App D. Obviously lots of thought went into it. And (even being a layering guy) I often think about Microsoft and wht their take (when they get around for forming one) will be on layering.

This feeling isnt new to me though. I remember having it in the WinFrame days, softricity, early VMware (think no virtual center just the web console) and of course now layering.

Your concerns are valid.

And I think there is a lot to the simplicity argument and people should always give consideration to simplicity (or complexity) of any design they are putting together. I also feel that you can pay for simplicity in someways. I mean setting a roaming profile attribute / just usign what MS gave you, and just having a romaing profile is much simpler than setting up a mgmt tool, with backends, and databases, and configuration and mgmt consoles. Just to manage and limit the profile. But simpler isnt exactly better in that case. Roaming profiles (left alone) can cause problems in the long run... thus why we wall pretty much hate them.

This is "some" of my point with layers and simplicity. I like the idea of a personalization layer. (hell we have a top level layer, two actually). It is simple just to have one on the top level. Though in keeping it simple I always sweat the model since there is nothing handle conflicts between user stuff (in a layer) and IT stuff (in a diff. technology).

Overall a well thought out article. And I can say you are MUCH MUCH easier to read when not pounding RES over the head with sledgehammer :-)

re: Does user virtualization matter? Goldman's $70m investment in AppSense says it might.

Kyle Joh — Mon, 08 Aug 2011 18:50:06 GMT

I have to voucher against Appsense because App virtualization is a double edge sword.

The product has it's merits and hassles but after more than a year of ownership. I think there are more negatives in the long run than pluses.

Some posters got it right, users will always have their way with IT and complexity will always lose in the long run.

The approach that Appsense uses for app virtualization has way too many holes to plug whenever an app changes in behavior through updates or work flow. The system breaks and Appsense Env fails to keep track of changes and users are without settings for days until we figure out what broke each time. This is more work than dealing with roaming profile management.

The best part is that you got archives of user profiles that you can restore back to the user. But the worst part is Appsense does not have enough tools for the admin to perform large scale rollbacks or delegate the restore to users or than someone with Admin abilities.

This is the only side I can say RES does better but their system is also complex and overly taxing on the system.

So, in short App environment virtualization is not there until Windows profile and app setting saving gets a complete overhaul by Microsoft. There's always going to be DLL hooks fighting and Appsense can't win all the time and forgets where it puts the settings.

re: Will layers matter in time to change the desktop management landscape?

Craig Miller — Wed, 04 May 2011 00:29:42 GMT

I think Microsoft is running around with tail between their legs in response to Apple's recent second coming with their iOS. Say what you will about Apple being company focusing on the consumers and not the business industry, but the fact of the matter is that IT industry is starting to become more and more like the consumer industry. If I were a betting man (and I'm not, I grew up in Nevada and I know how the casino's pay their power bills), Microsoft will be pushing all software vendors to not adopt the App-V, but to instead start adopting the new upcoming AppX platform which will be bundled into the core of next version of Windows. Of course AppX is just the next version of applications developed for Zune/Windows Phone Mobile that are packaged apps that will more then likely be directly hooked into the new Windows Marketplace, as windows becomes more like iOS and less like OSX. Of course I'm only talking about application layering here, but who is to say they won't try something layering profiles and other user data into the next platform...or maybe it's next one after that. Either way Microsoft has to do something because if they don't they'll shortly be just the company that hosts everyone's mail, because no one likes Exchange and many don't trust Google and their "Do no evil" demeanor.

re: Will layers matter in time to change the desktop management landscape?

appdetective — Mon, 02 May 2011 19:04:24 GMT

It's ironic that Citrix looses their CTO to do the user thing today. Thanks all for your comments.

I hope to see much more progress from the thinkers in this space. Clear the incumbents are going to do nothing, they are all too rich to take any risks.

re: Will layers matter in time to change the desktop management landscape?

Ron Oglesby — Thu, 28 Apr 2011 20:50:03 GMT

I wont talk Unidesk product here. Actually I think I go out of my way not to.

But general layering? MS would have been int he best position for this (as mentioned by Purnima). The problem here is that MS has done very little that is cutting edge on the desktop in a while. Like Softgrid (App-V, I just cant call it that) they have to go outside the house for new ideas and code.... Then they run SMACK INTO the mess that is any large silo'd business.

No one in MS wants to rock the boat on the desktop, and there is so much involved in making a change like this that it makes App Virt look like a simple service addition.

Anyway, I put my money on layering. Will it change everything? Not sure. You cant be sure of anything. I was at one time sure that MS would be virtualizing all their desktop apps by now and look how that turned out. Can layering REALLY help MS? yes. But only while being a threat to existing mgmt types. So the benefits have to be SO HUGE that they cant be ignored... kind of like Server virtualization early on.

re: Will layers matter in time to change the desktop management landscape?

Purnima Padmanabhan — Thu, 28 Apr 2011 19:02:21 GMT

@appdetective

Thanks for the insightful post. Few comments

1. On - Who can best do layering. No doubt in my mind that Microsoft is in the best position to do it. Short of that, we have one of the most mature layering solutions in the market. We see it as a critical step to getting to true single image management.

2. On – Will layers defocus us? We are first and foremost a management company and will continue to focus on and keep innovating in that realm. Today we are already hardware, os and hypervisor agnostic. Our goal is ultimately to also be delivery model agnostic. We want to be able to deliver a secure managed container for each person access to their data and apps irrespective or where it is running – on a client or a server. Today with the combination of type 2 and BareMetal we already support all end computers. Layers is the first step in extending that management to existing server based models.

3. On – Management of Type 1. We would love to manage a Type 1 hypervisor if there was one available with sufficient management hooks. Instead of waiting for that with bated breath for that – we are releasing our BareMetal – which installs directly on the host hardware. That said, as soon as there is a viable Type 1 solution, we will be one of the first to support and build management around it. Maybe Microsoft can up the ante here by delivering a solution.

re: Will layers matter in time to change the desktop management landscape?

Neil Spellings — Thu, 28 Apr 2011 16:50:28 GMT

Nice article.

I just love the fact there are RES VDX ads above and below the post.

Purely co-incidental, I'm sure :)

Shame @AppDetective isn't doing a BriForum Europe session - it would be fitting to have that Stig outfit at a UK event!

Cheers

Neil