“one size fits all” cloud for desktops is the wrong approach, I could agree more.

Well said. I like how this blog highlights different use cases of persistent and non-persistent desktops. I agree, let the end user decide what they want.

We rolled out a bunch of thin clients, and have a mixed environment of persistent & non-persistent Win 7 desktops....guess which have given us the most grief!? Whilst I think our deployment was rushed, the non-persistent users suffered losing their IE favourites, few things missed off the non-persistent templates etc. A lot less stress with a dedicated virtual PC!

Thank you guys for the validation. I do not think it is a big surprise that non-persistent was what people have done, given it really is just a derivative of session virtualization and persistent is just the next evolution.

Good post and largely agree,  
 
I would only emphasis that your point: 
"Let us build desktop clouds to meet the needs of the cloud users" 
 
should be: 
 
"Let us build desktop <s>clouds</s> to meet the needs of the <s>cloud</s> users" 
 
The key driver has to be the user experience and the cloud, VDI or client-side container (as we do it!!) is really just the form factor... 

Good point Stephen. I do wonder if in the end desktop clouds will be like the never ending battle like Windows and Mac and users will derive their own conclusion based on named solutions which have not showed their hands yet. So yeah, "Let us build desktops to meet the need of users which" which I believe belong primarily in the cloud :)

Whilst again i largely agree with the statements and argument presented the final decision comes down to "what best fits for the client?" persistent desktops work where users have or are allowed significant levels of personalisation including local deployed apps and non centralised data storage. 
 
In many organisations they will settle on a hybrid approach where task workers and power user have differing use profiles. 
 
The question has to be asked around what level of personalisation does this client permit on desktops as you can throw this around with the security team.  
 
Do you permit a user to put a phot of their kids/family on their desktop? what is the security around this image? there are many others that i have been discussing with clients and they all come back to dictated and implemented security policies. 
 
 
 
It all comes back to identifying the best fit for the clients use case and building a solid structure around the testing environment. 
 

Paul, 
 
I appreciate your insights into how desktop clouds have to ask the questions which deliver the answer to "what best fits for the client?".  
 
I would be interested from anyone out there what non-persistent can do in that regard that persistent would be different if you were using linked clone technology? Have any ideas there?

Peter, 
 
"Let us build desktops to meet the need of users" 
 
Isn't the answer to what is best for the company! If IT needs to discuss every configuration with every single user, we all keep busy for a while. Building end user environments is a delicate balancing act between freedom and IT department efficiency. To much user freedom and IT could be overwhelmed with compatibility issues, excessive testing, overheated malware scanners. You know like the rest of us it is never the shortcut to a document the users complain about. If life was only that simple! 

Cas, 
 
Great things to consider in your statements. I do believe freedom is not for every user hence my mention of meeting the needs of the user many of which freedom only inhibits their job. For those let us give them just enough. That might apply to most users even. 
 
Isn't the compatibility issues, excessive testing, overheated malware scanners something we already deal with with physical desktops but could be dealt with much easier (especially with new tools) in the cloud? 
 
Why cant life be that simple? Not yet perhaps, but on the path using persistent desktops...

Very well put. Totally agree with this. 1SFA (1 Size Fits All) is not very well recieved and apooses the trend of BYOB - which imo is a reaction to 1SFA and it-departments to slow to adopt to new technolgies and ways of use in the first place.  
 
Trying to lock these users down in new ways is imo not the way to go.  
 
Like that you state different use cases in the article. Good writing!

Gotta go ahead and disagree. The whole notion of virtualization is resources on demand. Your data, your apps, your persona, your corporate apps, an OS, etc. 
 
The technology now fully exists to deliver a workspace to a user using 100% non persistent means - user installed apps, customization, etc to such a degree they would not even know. 
 
While this discussion is healthy - i feel it is like arguing against dedupe because YOUR BLOCKS are actually shared with someone else, and arent really YOURS. Well, when you need them they are - and that's what counts :) 
 

Mr T.Rex, 
 
I appreciate your contrarian opinion and welcome further discussion, perhaps in a future blog. But there are things persistent virtual desktops do that non-persistent simply do not, like go offline for example.  
 
Shared is a good thing and linked clones with dedupe on storage still create what you are describing, however I fail to see why there wont be both and in the end the end user will decide.  
 
This all reminds me of the old Mac vs Windows discussions where IT delivers what they think works and in the end consumerization of IT prevails because knowledge workers get what they want.  
 
Let's put them side by side feature for feature and let them decide. 
 
Once last thought though.... What if IT isn't the same IT that is within an organization and instead is say a desktop cloud provider with all the bells and whistles bundling a desktop in the cloud with local endpoint management to deliver a best of breed xDaaS type solution?

I disagree, we're not making these assumptions based on the system making bad choices and not doing what you like. People are making these predictions assuming it works like it should. If we get to that point then yes it will be the future of our electronic experience without any question whatsoever. Of course assuming it works right also assumes it won't be a privacy, data, or user experience problem. Though we have yet to see anything like that conquered.

Erm why are we still talking about this ? This argument is not actually an argument, just a misunderstanding on the part of those who do not get non-persistency. 
 
You guys do not have an argument really and saying let the customer decide is really a sneaky way of pushing them towards persistency, especially if you dont understand non-persistency enough to properly explain it to them. 
 
OF COURSE they are going to choose persistent desktops, its what they know and are completely comfortable with.  
 
The whole notion of asking the customer to decide is just nutty and does no service to them whatsoever, are you seriously suggesting that when people come to you and ask your professional opinion that you give them two fundamentally different choices and ask them to choose from an uninformed perspective ? 
 
Just give them a non-persistent desktop, it does not have to be persistent underneath, the important thing is the illusion of persistency from the user perspective rather than the mechanics of the infrastructure. 
 
Seriously the argument of persistent vs non-persistent has long been won and the only reason to do persistent desktops is if you dont get non-persistency, lack the skillsets or tools required to handle non-persistency or you are just too damn lazy to bother building a virtualized infrastructure properly. 
 
OR if you have so many legacy incompatible (in the context of virtual) apps that its a complete nightmare and nobody has the balls to try. 
 
If non-persistency is cheaper from an infrastructure perspective, inherently more secure and easier to manage, why on earth would you be deploying persistent or even worse, asking the customer to choose. 
 
I have asked you this question before but have not yet gotten an answer from you guys, can you actually do non-persistency using your hardware ?  
 
Its just that I can think of no other reason why we are still rehashing this same old argument unless you guys just cannot do non-persistency, its difficult for me to understand and I am still asking why in an effort to understand. 
 
I am sorry Peter, letting the customer decide is just not a good enough reason and an easy way out of the question, like I said its a loaded question and if asked in the right way a customer will always choose persistency unless you properly explain to them that it doesn't matter and why. 
 
I kind of feel you guys deal with this subject badly, if you cannot do nonP thats fine but start working towards it because its clearly the way things are going. 
 
What you are doing badly is casting doubt on the whole P vs nonP question long after it has been answered and using the ignorance of the non-persistent model to support your assertions. 
 
Literally the only guys deploying persistent desktops are newcomers to the space who do not get it and hosted virtual desktop vendors stuck on an RDS environment and doomed to deploy persistent server slices disguised as desktops to their users because their infrastructures cannot handle non-persistency. 
 
I do understand it, its perceived as easier to do persistent, its less of an intellectual inconvenience and installing apps locally onto a persistent image is easier in the short term, its the way we have always done things, but it doesn't mean its right or best practice. 
 
Am going to have to call out this blog post as FUD on the basis you are blowing smoke around the issue. 
 
Locking users down ? Non-persistent desktops cannot do offline ? ALL FUD and bullshit. 
 
But feel free to carry on deploying persistent desktops and see how far it takes you, at some point you are going to have serious security concerns, grow yourself serious management issues as those persistent desktops bloat and outprice yourselves out of the market as non-persistent practicioners deliver consistently less expensive and efficient infrastructures. 
 
THEN when you do wake up and realise that the future of cloud desktops is actually the non-persistent model, you will be waking up at the same time as the rest of the late adopter market and have to scramble to enhance your nonP skillsets because you are just not going to have enough experience to properly balance application complex infrastructures. 
 
So we are clear : 
 
1) Its best practice to layer off your apps and virtualize them all, if you aren't doing this then you are doing it wrong. 
 
2) Its best practice to deliver a non-persistent architecture that provides users with the COMPLETE illusion of persistency. 
 
The art of non-persistency is delivering desktops that look, act and feel just like a normal persistent desktop without locking the user into anything and we have the tools to do this. 
 
T-Rex does not have a contrarian opinion, you guys do with your 'let the customer decided nonsense'. what T-Rex has is a best-practice mentality garnered from working at the top of the VDI game for close to a decade across a wide range of enterprise deployments. 
 
Brings me back to my question nicely, can V3 actually do non-persistency in a native fashion or do you lock your users into a persistent environment hence this whole blog post ? 
 

Brett,  
 
Thank you for your comments. I will echo your sentiment regarding  
"Of course assuming it works right also assumes it won't be a privacy, data, or user experience problem. Though we have yet to see anything like that conquered." 
 
I will add to that yet and V3 is committed with its partners to deliver on that. There are some interesting technologies in the works from many of our partners which pave the way forward where you can assume that is there. In the mean time the performance is a guarantee from V3 Systems regardless of which architecture a customer might pick. 
 
The data problem is one that has many answers and I think can also at this point be easily handled, however it will be handled in the cloud (public or private) even better in the very near future.

Guise, 
 
Wow! You are a passionate fella. I compliment you on your fervor. That being said, this argument against letting the customer decide is like a car lot that only has one model or make who has no choice but to argue that the customer shouldn't be making a decision based on choices but only knows one make of car.  
 
 
As the blog states (please read again) I am not making a decision for the customer and if you read a previous blog of mine, there are many architectures which could be offered with various use cases. I happen to prefer persistent virtual desktops for my own user experience.  
 
Now a question for you my friend: Why do you feel so strongly about something you only know one side of the argument about? If V3 really didn't want non-persistent then why invite you onto a webinar event knowing how you feel? How much do you really know about V3 and how our customers have decided? 
 
I realize it is disruptive to people that are in the architecture business competing for a given architecture, like say Citrix vs VMWare, etc to have to change and educate customers and let them decide use case by use case customer by customer how to handle which type of pool to use to fit their needs, but it is superior and its always been the advantage technology gives. 
 
I think it is worth discussing how a customer that employs multiple types of pools wins over those that think just one wins. Because a quote from just one customer illustrates this by saying "now that I have experienced this, I can never go back to the old way of just physical desktop management". 
 
To your first point, you might be right for some use cases, maybe even most, but what about the call center which just has one application? Why go through the added management, resource consumption for something which is a very simple instance of a desktop? Obviously there are many other cases which necessitate layering and other types of technologies which can apply to persistent and non-persistent alike. 
 
To your second, V3's customers have experienced otherwise. Have yours or you? Have you experienced a V3 experience? If not why are you arguing against change? I think we can all understand being contextually aware of the only architecture you are currently selling.  
 
Dan answered your final question again in LinkedIn and I will again here: Yes we employ many types of architectures on V3 appliances and do not require anyone to just use one type, specifically many of our customers have deployed floating pool on our appliances, perhaps even more then persistent.  
 
Much respect to T-Rex and yourself, however a simple question to ask about that 10 year tenure is have things evolved, and what are we evolving to?  
 
V3 and myself simply contend that just like previous technologies, until desktop cloud computing evolves to the point where end users choose experiences like desktops and dare I say it can migrate easily between them for their momentary purposes, then I have to agree with one Brian Madden who has argued many times that physical desktop solutions are superior. Since I do enjoy the V3 experience, I simply invite anyone to check V3 out and get an understanding of why he is out of context being that desktop clouds are now better then physical because they are a super set and simply deliver a customer experience unparalleled by one architecture solutions. 
 
Thank you guys again. I hope this was helpful.

Urgh, we are going in circles here and you are not adding anything to the conversation beyond simple analogies that lead back to non-technical buyers making seriously technical decisions that COULD impact upon them in a serious way. 
 
I am not just passionate and T-Rex is not just contrarian, we actually have a legitimate point of view that you are skipping around. 
 
Its really simple. 
 
If a non-persistent infrastructure can deliver the same 'persistent' desktops for less money whilst giving more security and making them easier to manage why o why would you use persistent ? 
 
Your car buyer analogy is flawed and is a different tack on the same 'let the buyer decide' argument, this is not a discussion that can be simplified so easily and its one that misinforms your customers in a negative way. 
 
The reason your customers choose persistency is because you dont understand non-persistency well enough to properly explain it to them. If you are giving the illusion of persistency to a user and from the user perspective nothing is different, then why stop to ask them which kind of underlying infrastructure they would prefer ? 
 
Its just lazy and easy to explain to your customers we take your traditional desktop, virtualize it and now it lives in the cloud, its CloudDesktop 1.0 thinking when we are at 3.0. 
 
In 3.0 the desktop is not even a desktop anymore, its just a pretty thing the app layer sits on top of and a policy engine, we gutted the desktop OS to cut down on the IOPS until its a shadow of its former self, a much more efficient way to place desktops into the cloud and this is not just me and my team, we see this emerging as best practice all around us. 
 
Remember the Windows desktop OS was never designed to be run in the cloud it needs significant hacking before it becomes something that we want to work with it running concurrently in large numbers on a server. 
 
Let the customer choose is a bullshit position, but if you feel comfortable letting uninformed non-technical buyers make a fundamental infrastructure decision in order to support your own perspective then brilliant, best of luck with that. 
 
I think the market will decide which companies die and which survive and I am perfectly happy with that, survival of the fittest is often the only way to distinguish between what works and what doesn't in a hype cluttered world of competing techs. 
 
I actually see this blog post as a direct response to the webcast I gave with Tal, it was clear from the way that you guys pulled it down immediately that you did not want to discuss non-persistency and now we see this blog post. 
 
Dan told me that he did not want to take it down but was overruled at the time by you guys and it did leave me feeling a little cold having it pulled so shortly after I took the time to appear and speak on your show, partly why I am bothering to spend time rebutting your 'let the customer choose' nonsense. 
 
You guys knee jerked and clearly did not like something Tal and I said (was it "there is no real reason to do persistent desktops anymore" ?) before dragging it down whilst leaving all the other webcasts up.  
 
Dan made a good show of explaining you were changing the site, but I understood him at the time. 
 
You only put the webcast back up when you felt you had a proper response to it, now you have and it's really no response at all that makes any sense. Until that webcast I had actually never seen you talk about or mention non-persistency once, am guessing this is a relatively new concept to you guys. 
 
I still do not get where you are coming from, I think you get away with deploying persistency because your customers are so small it doesn't matter, but when you get upto large numbers on single infrastructures then its a different game completely. 
 
I can only think that you push your customers towards persistency in order to take away some of the complexity involved and let them install native apps for their small businesses which kind of makes sense, it IS harder to virtualize an app than it is to natively install it. 
 
But I think you have dealt with this whole issue badly from the webcast to this blog post and I also still think that you do no service to newcomers to our space by casting doubt on the non-persistent model without actually giving any specific reasons beyond "our customers prefer it". 
 
Your customers prefer it because they do not know any better, if they actually did realize the significant security benefits of running non-persistency, they would not be using persistent desktops and if the buyers understood the significant cost-savings they would not be shelling out more money for persistent desktops. 
 
The answer is you are selling it wrong and not understanding the model properly OR you are just trying to sell more of your proprietary hardware. 
 
In answer to your question, I have never experienced a V3 experience and do not know what that is. 
 
I think you mean have you ever experienced VMware View running on SSD and if that is the case yes I have and I much prefer Citrix HDX, I find it delivers a better user experience than VMware can with PCoIP over the WAN. 
 
Thats what you guys do right ? You take a piece of hardware, brand it up to V3 systems, set it up to run VMware View and add a bit of your own proprietary software ? Is that a V3 experience ?  
 
So you guys sell VDI appliances to people who do not get VDI (if they did they would not need an appliance) and ask them to choose between persistent and non-persistent ? 
 
Is this correct ? If I am wrong please explain how, I wish to learn. 
 
In answer to your other question "Why do you feel so strongly about something you only know one side of the argument about? " 
 
I know and understand both sides of the argument, we were the first in the world to deploy cloud hosted desktops and back then all we did was persistent, its all anyone was doing back then. 
 
But we learnt along the way and helped the virtual desktop model evolve into its non-persistent form, we were the first to deploy these non-persistent desktops in the real world on scale. 
 
I work in the enterprise space and we would be dead if we were still deploying persistent desktops, which translates directly into "less security for more money" but I understand that you guys in the SMB space are still getting away with it. 
 
You ask what are we evolving into, we are trying to tell you. 
 
Worth noting that most of our users have no idea that they are using non-persistent desktops, they think they are using the same desktops they have always used. 
 
And YES you are making a decision for the customer with this whole 'let them choose' question. 
 
I repeat, if you ask a non-technical buyer to choose between persistent and non-persistent desktop without properly explaining it to them, then you are making their minds up for them in a sneaky way. 
 
Of COURSE they will choose what is most familiar to them, its human nature, but if they properly understand the benefits as explained to them by someone who properly understands the model, why would they choose persistent ? 
 
Finally I do have to pull you up on one point you made which illustrates more than any other that you dont get the non-persistent model. 
 
You ask "but what about the call center which just has one application? Why go through the added management, resource consumption for something which is a very simple instance of a desktop?". 
 
Non-Persistency is EXACTLY where you want to be doing with such a simple use case where task based workers are using just one application, in this case non-persistency would allow you to dramatically simplify management of those task user desktops and that one app.  
 
You ask why go through the added management and resource consumption, but this just underlines your lack of understanding of the non-persistent model. If you did get it then you would understand that LESS resources and LESS management is required to run these task worker desktops when compared to their persistent counterparts. 
 
Non-Persistency is all about doing more with less whilst making the desktops inherently more resilient to malware, if you dont get this then please stop talking about non-persistency like you do, its obvious you dont get it or like it. 
 
Am totally confused about your last paragraph which makes even less sense than anything that came before it. 
 
You said "until desktop cloud computing evolves to the point where end users choose experiences like desktops and dare I say it can migrate easily between them for their momentary purposes". 
 
I do not understand what you are trying to say here but I think its more FUD. 
 
I stand by my assertion that this whole blog post is FUD and hold it up as evidence that you guys dont get non-persistency or just want to sell more proprietary hardware. 
 
You seem smart enough to get it, why you refuse to I do not quite understand, I can only assume that your proprietary hardware/software mix doesn't quite like nonP or you just think that your customers will find it easier to install apps natively and to give their users each a dedicated persistent desktop. 
 
Who knows ?  
 
If a tech buyer is dumb enough to buy a proprietary VDI appliance rather than simply set up VDI on commodity hardware themselves then it makes sense to push persistency to them, it means you sell more of your proprietary hardware right ? 
 
 
 
 
 
 
 
 

I just wanted to add a disclaimer to anyone reading, there is a chance I am wrong, I often am !! 
 
I apologise Peter if I was a little harsh in my last post, its just that I have been arguing about MS licensing all morning with people and it tends to make me a little crazy.  
 
If you guys want to do persistent then thats fine by me, less competition for me over the longer term :)