Would You Move to Windows Thin Clients?

mcck asks: "My workplace is getting ready to study this problem in depth. From my preliminary research, there is very little savings when moving to a thin client environment that isn't based on Linux. Current thinking is that we will have to stick with Windows, so Linux is out for now. Citrix maxes out at around 10-25 users per server. I haven't studied Windows Terminal Server as deeply, but it looks to be about the same. Once we buy the 100 servers we would need to support 1000 or so of our users that would be migrated, plus increase our system administration staff to adequately support those servers, it looks doubtful that we would save much money at all, if any. Plus, if we upgrade all of the related desktop hardware to snazzy new official thin client boxes instead of trying to get more life out of old hardware (which is what they want to do right now), costs go up even more. So here's a question for anyone who has studied this issue, or seen its consequences at their workplace: Is thin client really a cost-saving approach to a large user environment?"

"Most users will be running basic MS Office apps, Groupwise for e-mail, and accessing some Oracle databases. A consultant hired for preliminary recommendations is saying that we should run Windows XP on the thin client boxes, not even the embedded version but the full one. Additionally, some of our users have more powerful applications like AutoCAD and ArcMap. We have already determined that those users will not be moving to the thin client machines.

Our department has spoken with a Citrix support/sales person who claims you can support up to 1000 clients on a single Citrix server. That seems so far from what I have generally read that I have a hard time buying it. Can anyone corroborate that claim? Again, most users will be using Office, Groupwise, and accessing Oracle DBs.

Does anyone have any experience with a workplace making this sort of migration? I would love to find a way to make it work, but from the research I have done so far, it doesn't look like we are going to get any cost-savings (unless they miraculously decide to go with Linux)."

Comments

A couple of considerations

[Future Man 3000]

Thin clients have less hardware to break. By their nature, it isn't likely a person using a thin client can mess it up from the software side such that a reboot won't fix the problem. The whole setup is more centralized, making it easier to address Windows patches and virus updates.

The savings would be better with Linux, but they may very well be worthwhile anyway. Determine how much IT time you're going to save against the cost of the setup.

Re:A couple of considerations

[narrowhouse]

A REAL thin client has less hardware, and that would be good and possibly a savings over time. If they are running a full version of XP on each unit as mentioned in the original post they are just using PC's as clients (by the way ask the consultant who told you this was the way to go how you are going to save money by A) Buying full versions of XP for each machine. B) Terminal servers C) AND a seat license for each computer/terminal, if the answer isn't "well er... uh... I guess you won't." there is a good chance he is either a liar, idiot, or Microsoft rep.)

Re:A couple of considerations

[llefler]

There's more to it than that if you truly centralize. Assuming 40 users per server you're looking at maintaining 1 machine instead of 40. That machine is in a controlled environment. You only have to patch the machines in the computer room. When you need to upgrade software you only have to do it once. You don't have to deal with users installing software, or gator, or the latest and greatest Outlook plugin. You should have identical servers, that way you can create an image and build a complete replacement in minutes. (I think 25-30 servers for 1000 users would be reasonable) If you look back at our past a little, it wasn't uncommon for one IT person to support 200-300 mainframe users. So one payoff is a smaller helpdesk staff.

Then if you replace those desktop PCs with something like a WYSE terminal, you end up with something on the desktop that is probably good for 10 years. Compared to the disservice you're giving your users now by making them use those 5 year old Win 95 machines.

Now of course the idea of putting Win XP on existing PCs as thin clients has to be the most ridiculous idea I have ever heard. I suppose if those old machines aren't large enough to handle it, you're supposed to buy new PCs to replace them.... If they're going to go this route, install the terminal services client on the existing PCs. They should start moving their core applications to the TS servers and removing them from the desktop. Then as they replace systems, put non-power users on dedicated terminals. Shoot for balancing costs by reducing staff or freeing them up for other support tasks. And then look for savings in desktop hardware costs by increasing the useful life of the equipment.

I'm curious to see where this trend goes. While X-Terminals and Linux servers would be great. RDP terminals and Terminal Services will work for those 'gotta be MS' bosses.

Careful how you handle your users though. No local CD means they can't use 'their' PC to listen to music. They also won't be able to copy files (that they probably shouldn't copy anyway) to take home and work. Part of centralization is IT taking back control of the computing environment. And some users aren't going to like that.

Re:A couple of considerations

[dpoulson]

You have got the option for the thin client to run linux (maybe a custom knoppix) with rdesktop which will connect to windows terminal server. If you use Citrix, of course there is the citrix client for linux.
Both of these options mean that whilst everyone would still be running windows as far as they are concerned, you could save a lot of money on the thin client front. Maybe even re-using your old PC's (always a good option to tell the bean counters!)

Re:A couple of considerations

[Glonoinha]

-The real payoff is with the centralization.

Bingo! At a company I did IS/IT at quite a few years ago they (hmm as there were only two IT guys I guess I could say 'we') would buy random clone machines and switch gray market vendors to shave $50 off the price of a $1200 486DX/2-66 box. The proliferation of different hardware brands, interfaces, etc. was a tech support nightmare and we ended up needing to hire another IT guy just to keep everything running.

The minute you need one more guy to support your infrastructure it costs you $50,000 plus benefits and HR overhead to make up all the times you save $50 on the price of a machine. Ouch.

If you (through centralization) can support all those end users with the same number of IT staff, instead of adding ten more guys, over the course of three years (lifecycle of an operating system, machine, etc.) you are looking at a savings of $1,500 per desktop.

Looking at it as well

[mrscott]

I've been looking at this as well but on a smaller scale. First, Citrix doesn't support 10-25 users per server -- it's per processor. So a 4 way processor can conceivably handle 40-100 users and since you don't seem to be running hard core apps like CAD (you mentioned that these users would not be migrating), etc, let's say 20 users per processor or 80 users on a 4 way box.

That puts you at 12 Citrix servers.

Next, according to the Citrix folks I've worked with, Windows Server 2003 handles Citrix MUCH more efficiently than Windows 2000 resulting in -- according to them -- a doubling of the number of users possible on each server. Since I don't quite buy that, let's go with a factor of 1.5 times the users rather than doubling. But let's stick with 12 4-way Citrix server to account for redundancy which you will surely want for this solution. Heck -- let's go with 15 even. It's still a lot fewer than 100.

I agree that -- at the beginning, using old hardware would save money initially. But, consider the support angle for a second. Rather than new, identical thin clients, you'll still have whatever you currently support for desktops. If you just maintain the copy of Windows that's on there, when one breaks, it's a total reload. If you use the new thin clients, it's a matter of swapping out the unit and they're less expensive to buy initially as well as more reliable due to fewer moving parts. You should see support costs drop dramatically with this rollout.

The Citrix guy that mentioned 1000 users on a single box had to be talking about something much larger than 4 processors... personally, I would recommend a cluster of 4 ways servers for teh redundancy that it would provide in the event of a hardware failure.

Going with Linux won't necessarily save in the long run. Sure -- you'll save on the initial software acquisition. But consider the support, end-user retraining and other problems that could crop up. Even if you use Cross Over office or something like that, your users WILL require retraining and you will suffer a productivity issue initally.

Not going with a Linux solution won't automatically doom the project nor will it prevent savings from the implementation.

I'm all for Linux (I have it deployed where it makes sense), but am wary of making generalizations that it automatically saves money.

Re:Looking at it as well

[llefler]

Multi-processor doesn't scale that well. 4 processors doesn't mean you can support 4 times as many users. Although it will probably mean more than 4 times the cost. Use 1-2 processor boxes. (we have racks of Netfinity x330 servers) Use Terminal Services load balancing or put a hardware balancer in front of them. Fewer eggs in each basket and cheaper too.

And our experience with Citrix is that it adds a whole lot of $$$ to the equation without bringing much to the table. If you aren't using non-RDP clients, I'd pass on Citrix.

Oh, and don't jump on Win 2003 until you know that your apps are going to run on it.

6 vs. half-dozen

[smoon]

Other comments have suggested 10-20 servers is a more realistic number of servers, but server hardware tends to be a _lot_ more costly than desktop hardware, so 20 servers at $5k each = $100k, vs. 100 desktops at $1k each. Plus you have to license citrix, windows for every client, buy at least a few new computers, or get some kind of 'winterm' (about the same cost as a low-end PC -- $500-700 with monitor). Don't forget the implementation, training admins, hiring or contracting skilled WTS/Citrix admins, upgrading network infrastructure, etc.

Based on the above, lets call initial costs a wash, or perhaps thin client costs twice as much compared to normal PCs.

Now it comes down to whether you believe the gartner TCO b*llsh%t. Make some wild-assed assumptions and you can come up with a 'cost model' that will support any position you want to take.

Our experience:
Citrix/WTS works well for remote access and for running apps from non-windows platforms. It's also good for remote offices where the function is well defined and there's a small number of users (e.g.: retail locations).

Problem: Security is essentially non-existant. There is a veneer of what appears to be a 'locked down' environment, but there is always some way to open a help window or somesuch and pop into IE and then *bam* you've got the ability to browse the file system and run just about anything.

Another problem: Because Citrix/WTS allocates memory per user and does not share between users, vast amounts of memory are needed per server. We have an app that uses 180MB per user, so for 10 users we need around 2GB of RAM. This puts a hard upper limit on how many users can use a memory-intensive application.

We also use Linux and X-Windows. Running a java swing app that takes about 80MB per user, around 70MB is shared, so each user takes a delta of 10MB. That same 10 user mix uses around 170-180MB of RAM vs. 800MB+ if it ran on Citrix. Also, since much of the app is shared, CPU cache is much more effectively utilized and performance is much better as well.

Perhaps you could concentrate on migrating apps to X-Windows and running them on Linux servers. This won't work for everything, but it may end up providing a lot more business value.

Good luck. You'll need it.

Fire your consultant ...

[molarmass192]

A consultant hired for preliminary recommendations is saying that we should run Windows XP on the thin client boxes, not even the embedded version but the full one.

Your consultant is an idiot, fire him. Besides that, you have to decide if you're going to seriously persue a thin client strategy or not. If you're doing this for cost savings and don't want to move to Linux on the client side, look into VNC or TightVNC. However, this is a not a real thin client and to be honest I wouldn't do it becuase it's probably going to fail.

If this was my shop and we were serious about saving money, I'd go with Linux or BSD on the existing client side hardware and run an X server based thin client. If your hardware is a fairly standard configuration, you could even go with a diskless setup and remove all admin needs on the client side. The windows application server could be implemented using VNC running on W2K -or- Win4Lin terminal server on Linux -or- Crossover Office Server on Linux. This won't allow your CAD user to migrate however, unless a Linux flavor of their software exists, they're stuck on Win.

Back to your consultant, the key to saving money with thin clients is to reduce costs on the client side. Springing for XP and hardware upgrades on the client is 180 degrees from your stated goals. My guess is he's merely reciting his MCSE crap to ensure he can come back for future billable work.