Desktop as a service

[Ted Giles]

See here.
Whats this all about?
http://go.techtarget.com/r/22102397/6609138/1

[mronck]

Re: Desktop as a service

They pretty much explain it in that very same article. What's your question about this Ted?

The idea of a virtualized desktop is to make better use of resources. When you have 10 desktops, most of the time they will not be used all at the same time, or not all in the same way, or not all processes that run will demand the same resources. As a result, in your 10 PC Desktop parc, maybe 7 are really necessary could you balance the resources from those 7 PC's more optimal. By virtualization you can. Hence, it might very well be possible, that through virtualization you only need resources equal to 7 desktop PC's. That results in savings directly related to resources. Savings=Money.
Besides that, there is better security (since there are not 10 PC's with data in 10 Rooms but all are in the same server room) and better maintainability. Furthermore, you don't need fat desktops, but can do with thin clients which are much cheaper.

Once you decide for a virtualized desktop system, you can ask yourself if you want to do it yourself, or outsource it to an external service provider, in which case they call it DAAS.

[Ted Giles]

Re: Desktop as a service

Question is, what real difference is there between this and thin client?
More marketing hype and a new name for an old principle?

[mronck]

Re: Desktop as a service

Well, sometimes the venom is in the detail, not? A thin client in itself is just a lean machine. Furthermore, a thin client is a PC. Hardware.
Virtualization techniques are software. A desktop can be a fat client or a thin client. Hence, you can virtualize both a fat client AND a thin client.
If a thin client is virtualized, what I wrote in my first response comes into effect. By now there is much information about virtualization available on the internet. Just do a search and many articles will pop up.
VMWare is a good starting point to see what has been accomplished until now.

[DaveM]

Re: Desktop as a service

desktop as a service is very workable for some. I would definitely go there if the costs were in line and I had a locked application where no one was supposed to do anything else.

But when you can get a computer for 499 at best buy or walmart that will run the shadow, one would have to look at the costs over all. Most of my clients do this because they are not just running one application.

Just weigh those costs and needs against a thin client where the computer cost is??? and costs of the extra memory on whatever server???

[mronck]

Re: Desktop as a service

That may actually surprise you, since you need way less PC's then you would need as standalone onsite hardware. Which does not only mean the price of the PC, but also OS, Security etc etc.

[DaveM]

Re: Desktop as a service

That may actually surprise you, since you need way less PC's then you would need as standalone onsite hardware. Which does not only mean the price of the PC, but also OS, Security etc etc.

I can understand security, but os is included with the 499 computer and if I have 20 users, I need 20 stations anyhow. What is the cost of a thin client computer type hardware? I am seeing 350 thin client with a monitor, keyboard and mouse for thin client and 499 for a desktop.

I think each person has to weigh the need, cost and security.?!

[mronck]

Re: Desktop as a service

Yes, of course, each person has to weigh need, cost, and other parameters to the investment. That's common business practice, nobody just throws money out of the window. Experience learns however, that in cases where the desktop as a service comes to effect, we are not speaking about 20 users but mostly about hundreds of users. Figures tend to change quickly in the advantage of the DAAS solution if numbers go up. Figures go much further as just the comparison on raw purchase price alone. If you would need to maintain say 200 desktop computers you would need 2 guys on your Support Department to do it: answer user questions, solve small daily user problems, install updates, initial softwares, repair hardware, you name it. There can be a huge saving on personnel cost besides other savings.
Scalability is only one issue, as you see there are others that should be implemented in a calculation. Virtualization has many advantages however not only cost savings. Security issues are also important and in many cases make the final decision. Distributing a new software is a matter of deploying an image overnight and applications availability/business continuity can be secured much better. Dynamic resource allocation can give an extra boost to workstations who are in need of it due to a heavy process running which improves performance. Fat clients are hung-up to their hardware specs, thins are not, desktops can be configured outside the thin clients resources due to resource reshuffling on the servers. Connectivity can be much improved due to options for carrier independent connections to the backbone at the providers data centre. It would be very costly to get that connectivity in your own office. Same goes for power supply and other bc-configurations. It often is a complex package when you want to calculate the advantages of something like DAAS. Same goes for any other virtualized solution. However, in some institutions policies go above cost for instance because of federal/government regulations, Laws applicable or demands from certification processes companies take part in.

[TPeterson]

Re: Desktop as a service

Marcel, our IT guy decided to implement a virtual desktop environment in our 25 person company about 12-18 months ago, and so far it's been a nightmare. The server(s) seem to go down weekly, which means almost ALL work stops, and people complain about speed and other issues very frequently (personally, I've refused to be part of it). I remember our IT person explaining all the benefits to me when he decided to go down this path, and of course, it seemed to make perfect sense. In hindsight, as you say, this is meant for larger installations of hundreds of users where REAL savings can be realized. Also, we have since learned, via a third party consultant, that our investment in virtualization was insufficient...I guess in relation to having the necessary built-in redundancies (I've not been party to those discussions).

Anyway, this is not meant to contradict anything you've said regarding it's advantages when done correctly, with the appropriate investment and under the right circumstances.

[Bill Griffin]

Re: Desktop as a service

Data centers use vmware regularly. It is a rocksolid platform that (knock on wood) we have never had go down. We use VMWare in house and I could not be happier. There are ways of building in redundancy that if a server goes down, you should not even notice a problem. Setting up redundant servers are not inexpensive, but well worth it if you need to be up and running 24/7.....I am not sure what your IT guy tried to do or what he was using as a platform, but Vmware is a proven, solid system.

[TPeterson]

Re: Desktop as a service

Bill, I'm not sure where the problem lies with our installation. I would guess that our IT person tried to implement something not TOO far above his normal budget in order to get it through and is paying the price now. He's a pretty knowledgeable and thorough person, but there could have been some elemental issue(s) overlooked. This could be another case of vendors selling a solution, where a solution really wasn't needed, and not pricing it out appropriately.

[mronck]

Re: Desktop as a service

Marcel, our IT guy decided to implement a virtual desktop environment in our 25 person company about 12-18 months ago, and so far it's been a nightmare. The server(s) seem to go down weekly, which means almost ALL work stops, and people complain about speed and other issues very frequently (personally, I've refused to be part of it). I remember our IT person explaining all the benefits to me when he decided to go down this path, and of course, it seemed to make perfect sense. In hindsight, as you say, this is meant for larger installations of hundreds of users where REAL savings can be realized. Also, we have since learned, via a third party consultant, that our investment in virtualization was insufficient...I guess in relation to having the necessary built-in redundancies (I've not been party to those discussions).

Anyway, this is not meant to contradict anything you've said regarding it's advantages when done correctly, with the appropriate investment and under the right circumstances.

I am sorry to hear about your experiences Tom. But these experiences are quite common I must say. There is A LOT that can go wrong in the process of selecting what kind of services you purchasing from whom exactly and what infrastructure will be made available to you. That does not say (as you pointed out) the choice for DAAS was wrong, but rather that the choices made were not optimal. This can be because of a lack of knowledge, lack of budget, or simply the wrong choice of provider. I always saw the provider as a partner that really needs to be on your page and side. In the process, they are very welcome to make some money out of it, but first objective needs to be always to meet our requirements. Speed is a matter of resources and connections. Nothing else. But both require money and the right designs in set up.
As you also said, investment in virtualization needs to be sufficient to meet the requirements. I would say that if you need to be on a tight budget, this might not be the wisest choice. I would not perse say, that 25 FTE is to small a basis, it all depends on what your quality requirements are. It seems, the pre-quote work has not been done right.