BriForum 2015 London kicks off today! (If you're not here at the show, follow the action on Twitter and Flickr.) One of the cool things we started doing last year is to ask attendees when they signed up what question they'd like to ask me. So here's the list of all those questions, plus my answers. I actually wrote this straight through without editing or stopping. So these are my real unfiltered thoughts.
All the parts that appear to make up a desktop virt solution seem to require a large amount of different vendors. e.g. Citrix for the protocol, AppSense for UEM, FSLogix for app layers. Is there a single vendor solution that scales well?
I don’t think the single vendor solution thing will ever happen. The problem is that VDI & desktop virt are really all about Windows desktops and Windows applications, and those are very diverse and complex. So just like there will never be a single product to manage, secure, deliver, update, etc. your traditional Windows desktop environment, I don’t think they’ll ever be one for your virtual desktop environment either.
As a fellow blogger and writer about Citrix products. How do you find the time to stay on top of everything you blog about??
I don’t have any users. :)
Can you ever see Desktops as a Service or VDI moving into to the consumer home market and being adopted in people homes in place of the traditional desktops we see today?
No. DaaS / VDI are about Windows apps. Normal consumers don’t need Windows apps. Sure, they use Windows apps if they have Windows, but they use Mac apps if they have Macs. And most of the “apps” people use today are actually web-based services (Email, Facebook, etc.) I just don’t see the value in a consumer getting a remote Windows-based desktop. (I can see value for Chrome OS and other platforms like those, and I could envision those being remotely managed.)
Can you see a day when ALL desktops will be thin clients or run a device that runs a simple OS.
No. We keep on adding too many new types of apps and devices. I really feel like if remote Windows apps and thin clients haven’t become “everything” by now (after 20-ish years), they will never be. (That’s ok though. We can still use them where they make sense, and we can still focus on better management of traditional desktops, but I just don’t see all desktops becoming thin clients. Heck, even iOS is not a true “thin client” as there are lots of apps, data, and user personality installed locally on it.)
Convergence...(PC to Mobile)...when and in what form do you see this meeting?
For non-Windows apps, we’re seeing it today.
I like VMI (or some version of Windows 10 mobile / Android apps delivered via RemoteFX / HDX / PCoIP / whatever to mobile devices, and I like app refactoring (PowWow, Reddo, etc.) as a way to make traditional Windows desktop apps usable from small-form factor devices, but my core belief is that the type of things you do when you sit down with a huge screen, keyboard, and mouse are fundamentally different than the types of things you do from a mobile device, so it will not be possible to have the full convergence of every app on every device.
Cutting through the vendor hype, in your opinion, what would be the one area you would address to ensure that user experience (performance) did not become an issue.
Two unrelated things. (But since you said one I’ll call them 1A and 1B :)
(1A) Put GPUs in your VDI servers. It only costs 10% more in terms of your capital outlay, and it solves so many problems (in terms of compatibility, user experience, etc.
(1B) Use storage that can give your users hundreds (I mean literally hundreds.. 200, 500, 1000, even 2000) IOPS per desktop. The stat I heard the other day was that a Windows desktops needs 2000 IOPS at 2ms latency for the disk *not* to be the bottleneck. Modern storage solutions can support this.
Deploying and maintaining an on-premises VDI environment is hard, expensive and a bit boring. Can I just pay Microsoft, VMware or AWS to run it for me now (DaaS)?
Easier said than done. More details in this article I wrote a few months ago.
Discuss the merits of using an iPad as an enterprise thin client, docked to a full size keyboard, screen and mouse, running Citrix applications, browser applications and native apps. Is it an emerging set of technologies or are vested interests currently preventing it happening?
Problem I have is I can’t multitask with local iOS apps, so now my iPad is a $500 thin client, and not a very good one at that. (No choice of mouse, must use Bluetooth keyboard, only one display, so-so performance.) Why not just buy a $200 thin client and let the user use the iPad how they want?
Put another way, what is the advantage of using an iPad as an enterprise thin client in a dock? Seems like an expensive hassle. An iPad is meant to be used while people are walking around. Where are all these docks going to be? How much do all the connectors and stuff cost? Why not just buy normal thin clients?
(This is kind of related to the Nirvana Phone concept Citrix’s Chris Fleck talks about that I’ve addressed before. We actually shot a video with him at Synergy last week that we’ll publish this week. I’m still not sold on the concept. Neat party trick though!)
Have you ever heard of Awingu?
I would ask him why he is not listening to Benny. Because I think RDSH is still the better solution ;-).
Check out the video of my session from Synergy last week called “RDSH versus VDI: 2015 Edition.” I still like RDSH and agree there are many cases where it is better than VDI. And I can’t always listen to Benny because he talks a lot. (And he says the same thing about me I’m sure. :)
I can’t figure out how to direct link to the video of my session. For now go to this site, click on May 13, and scroll down to the 11:00am - 11:45am section. My session is SYN318: Terminal Server versus VDI: 2015 edition.
In a smaller organisation with a limited budget and many remote locations, what would you say is the biggest disadvantage of VDI?
You should only use VDI when you need it. In other words, only for the scenarios where you need remote Windows desktops and applications, and only for the subset of those scenarios where multi-user / session-per-user / RDSH won’t work. Taking that, if you need VDI, then you need VDI. Period.
The biggest disadvantage of VDI is that it’s complex, expensive, does not provide a user experience that’s as good as local, and it only works when the user is online. Of course there are advantages too, like the fact that user can logon from anyway, they can “flow” their sessions from device to device, and (in some cases), it’s easier to manage and easier to secure. (Those are not automatic, but they’re possible.)
For lots of remote locations with a limited budget, do you need VDI? Why? What are you trying to solve? Can just publishing some apps with XenApp work instead? Lots of questions . . . VDI is certainly not the automatic answer (though it could be the answer).
It seems there's still a place for reverse seamless technologies to provide end users with the full desktop experience they demand. Will this ever be made redundant and how close are we?
Hmmmm… Reverse seamless is awesome in concept and tough in practice. My main question about it is if users need an app or two locally, why are you delivering a full remote desktop? Why can’t you just deliver seamless published apps (even if it’s persistent VDI) and then let them access them and the local apps all via their local desktop?
I do not see reverse seamless taking off. Maybe if it was really there 10 years ago, but not today.
Software layering solves some inherent VDI issues. What would be its value in non-VDI scenarios (RDSH) and with regard to the need for App-V (labor intensive as that is)? Published apps is a trend for servicing unmanaged (mobile) devices with different form factors. How does that affect the balance of VDI/RDSH in favor of the latter?
I don’t think the two are related. Software layering (or whatever you want to call it. App-V, Ceedo, FSLogix, App Volumes, Liquidware Labs, Unidesk... etc.) is becoming more compatible with more apps, and the overhead “tax” of using it is not as high as it used to be. This just means that you can get more apps on fewer base images, regardless of whether it’s VDI, RDSH, or even using the layering solution on your traditional desktops. So I don’t think it’s related to the VDI-to-RDSH balance really.
What’s the future of Citrix XenApp?
Citrix Workspace Cloud. (Seriously!) More details in this article I wrote a year ago (from when Citrix Workspace Cloud was called Citrix Workspace Services).
Also, unrelated, Jack Madden and Gabe Knuth are doing a session here at BriForum about VMI (Virtual Mobile Infrastructure), and I like the idea of a future version of XenApp being able to deliver remote Android and/or Windows 10 mobile apps to mobile devices via HDX, just like how it delivers Windows desktops apps today.
What do you think is the biggest missing feature in application streaming solutions known today?
I just sent an email to Tim Mangan to ask this. I’m not an expert on app streaming / app virtualization, so I don’t have a good answer. I’ll paste in his answer when I hear back.
Answer from Tim: "Doing 100% of the apps. There are a bunch of little things that can also be done better, but getting the remaining apps manageable is the most important. And while I expect that situation to improve, it is unlikely that vendors ever get to 100%."
What do you think is the most complete UEM solution in the market? Why?
It’s hard to say since “complete” means a different thing to different people. They all have features that are unique to them, and they all seem to be missing other features. I will say for sure it doesn’t come from Citrix, VMware, or Microsoft. :)
What industry event has caught you by surprise since you started BriForum and what hasn't happened that you really thought would?
Like almost everyone, I was surprised that mobile is such a huge thing (both with phones and iPads). In fact when Jack started with us full time in 2011 or 2012 (I forget when), we sort of gave him mobile because he was the “new guy” and Gabe and I didn’t want to deal with it because we thought it was some crazy thing for the young guys!
I’ll have to watch my videos from those early BriForums when I have time. I know for sure I thought that we’d be streaming Windows disk images from the cloud by now. That is way wrong.
I also thought that lots of these old Windows desktop apps would be converted to modern or web apps. That also is not happening. (And now I believe that Windows apps will be around forever when it comes to the enterprise.)
I’m also kind of surprised that the full Citrix suite still has a million different consoles, and that it’s a “suite” in name only, just like 2003.
What is your favorite product that was released this year (2015)? Or which of the new released products (2015) do you recommends (most)? (Also: What is the hottest virtualization product/concept the coming year?) <— I combined this questions since they’re similar.)
New app layering / app management. FSLogix, Unidesk, VMware App Volumes, Ceedo, etc.
Also I like the hyper-converged products that let you use your own hardware (Atlantis and many others . . .)
Finally I think the products that allow you to refactor Windows apps for mobile users are awesome (PowWow, Reddo . . .)
Oh, I love that VMware entered the RDSH / published apps space because it finally lit a fire under Citrix who was getting complacent. (I think Citrix’s extension of XenApp 6.5 support to 2017 is an example of a direct result of pressure from VMware, since if people have to move off of XenApp 6.5 and XenApp 7.x is not fully featured, then why not shop around a bit??)
What is the biggest mistake you can make when implementing a VDI solution?
Treating VDI users different from traditional desktop users.
Also trying to do too much at once (i.e. migrating from physical to VDI at the same time as also trying to go from the “Wild West” any user can do whatever they want persistent desktops to locked down non-persistent desktops.)
Using VDI when you don’t really need it. i.e. When RDSH will do, or when you just want “better management,” but you get that better management by moving to locked down non-persistent desktops which is something you don’t need VDI to do. (You can just use Ghost and group policy to do that on your physical desktops today.
What is the next big thing to worry about? We've done server virt, desktop virt and app virt... What's next?
Trying to figure out if you need a mobility strategy and/or if you really need EMM. Also trying to figure out how to get your enterprise Windows desktop apps onto mobile devices. (Remote apps? App refactoring? Rewrite the apps? Web apps? Something else??)
What virtualisation software does Tony Stark use?
A proprietary solution he wrote it himself. (Not even QEMU-based!)
What would you say is the one thing you got massively wrong. I'm thinking techonology trends but feel free to share.
That one’s easy.
When will Microsoft change how they license VDI environments? (When they provide DaaS themselves...)
“When the provide DaaS themselves.” <— Bingo!
Which reasons can push or drive a VDI solution in a manufacturing industry? What is your approach to be succesfull to transform a standard physical environment into a virtual one?
I don’t know about manufacturing specifically.. I think all the things that drive standard reasons for VDI, which are (1), you have a good reason to need datacenter-hosted apps, AND (2) you have a good reason not to use RDSH (or an RDSH solution like XenApp, Horizon 6, vWorkspace.)
As for the approach to be successful to transform a physical environment to be a virtual one, to me the most important thing is to not try to do too much at once. For example, if your physical environment is persistent desktops, then do not try to change to non-persistent and go to VDI at the same time. That is too much to bite all all at once. Instead go to non-persistent /locked down physical first, get that working, then go to virtual. Or do a P2V of your persistent physical environment to VDI first, get that working, then start working on moving to non-persistent.
With Microsoft Windows 2012 R2 and the upcoming release Windows Server (10?), how relevant is Citrix XenApp against Microsoft RDS only
Same as it’s always been. Better protocol, more clients, more complete “package” that’s enterprise ready right out of the box, better management, etc. This is Citrix’s core job, not Microsoft’s, and Citrix (and VMware with Horizon 6) are more focused on a better end-to-end solution than Microsoft.
With so many different options and solutions for storage now available, is Microsoft's VDA license requirement for non-Windows based thin clients the last barrier to widespread VDI adoption?
No. Microsoft licensing sucks, but even if Microsoft made VDI “free” (just with a base Windows license and no other shenanigans), VDI is not going to be unleashed overnight.
VDI is “not” the future of the Windows desktop. It’s good for specific cases where you need the advantages of a Windows desktop or Windows apps in the datacenter and where RDSH won’t work, but I still believe that’s only appropriate for a small percentage of your users and/or your apps.
I would to like to understand the best suitable option for our application virtualisation. How can we define requirements for an application to choose either DaaS or SaaS/PaaS?
We wrote a whole book about this!
I’m assuming you’re referring to virtualizing the app by putting it in a DaaS/SaaS/PaaS environment (XenApp, etc.) and then delivering it remotely?
If you want to look on an app-by-app basis, you need to look at “what else” the app uses. Where’s the app’s data. What file and databases does it need? How hard is the app to install. (Can it be easily packaged with App-V, etc.?) Does the app work well remotely? Is a Windows client the only option, or is there a web app?
That’s it for this year! Feel free to hit be up here at BriForum this week if you want to talk further, or share your thoughts in the comment section below. Happy BriForum!