Home Forums Thin vs Thick Client Comparison

Interesting read: http://www.knowledgeonecorp.com/news/pdfs/Thin%20client%20vs%20Fat%20client%20Computing.pdf

Perhaps sticking with desktops is a better idea. What do you think?

While the author of the quoted paper has a right to have is opinion, I do not agree to his opinion for the following reasons
1. Almost all client server applications store data in the server
2. Server side scripting dos the major job of computing. It needs resources any how
3. Thin client (No CD, hdd, USB etc) has lower maintenance need. This also improves security.
4.Administrators need to concentrate only on server. Software updates/ Patching restricted to server.
5.One good example is the Indian Rail Reservation system which perhaps the largest passenger reservation system works on it. This is in itself is a proof is utility.

I do realize there are benefits to thin clients, just not when there's many applications or medium to heavy resource usage involved.

1. No argument was made about where data is stored. It's all about processing power. Your new server will have to be able to process the same amount of data as the group of desktops you had before.
2. Server side scripting does the major of computing in certain applications. Any sort of graphic-intensive workload is not possible. Say for instance, any modern game requires so much graphics processing that you would need one crazy server to render the environment. If you're processing a lot of high resolution images like those in medical imaging, you may run into problems. Embedded videos like those on youtube and such can cause a dramatic slow down as well.
3. No USB can be a big negative. Especially considering how many people use document scanners (ScanSnap). This would essentially render them useless. Anything that requires specific drivers can't be used. A lot of practices still burn patient charts/files so that no CD thing can be a negative as well.
4. If anything happens to that server, your entire operation is down. At least if something happens to a regular server, the local applications are still there. Just because it's all on one server doesn't mean those issues with software will disappear. Honestly, how often is it the hardware at the local workstation vs what the person did on the workstation? Those issues with Office and other applications will still persist. Furthermore, if person A opens an email (with a virus) or a downloads a a virus on the terminal server, no matter how impenetrable you make it from the outside, the entire server and every user is compromised.
5. I can name about 500 other companies that use workstations. Most of them in the Fortune 500 . That reservation system involves a proprietary application, CMS. They are using thin clients for nothing else other than the management system. In a real office, people are bouncing between Microsoft Office, emails, SQL applications, browsers, videos, pictures, etc.

WSUS and USB caught my eye. I just think there is so much more versatility with regular workstations.

I'm surprised not much has changed from when this paper was published in 2002.

So Brock, have you been getting my messages on YouTube? What have you decided?

Got to jump in.
I think Thick and Thin clients each have their respective places, without a once size fits all, AND to boot, there are thin clients that DO have USB ports.
We (my practice, and the 500 bed comm hosp of which I am CMIO) use a mixture of W/S (thick clients) and thin clients.
For the most part, thin clients are easier to manage from an ISS standpoint, and thick more useful for the technically ept (as opposed to technically inept) user.
So now that I've jumped in, I haven't added a blasted thing to the thread.

Those thin clients that do have USB ports are usually for generic devices. Like a flash drive or a keyboard. Most don't support specialized devices like scanners or non-generic printer drivers.

Sandeep,

If I have a server in my personal office serving as the desktop would that be considered thin or thick? and would you see an advantage to that scenario or better of with a desktop/server as my main computer in my personal office? maybe my question is too basic, but I feel like a first grader talking to a grad student with this stuff.

That would be neither a thick nor thin client. That's the actual server. That's not really what we're getting at here. It is strongly recommended that you do not use your server/main computer as a workstation/desktop. It's best to have a dedicated server/computer for AC. The argument about thin vs thin client deals with whether or you have full-powered desktops or thin clients that connect to your server. Thin clients essentially remote into your server. All processing is done on the server itself. Thick clients grab the data from the server and process it. To use thin clients you really need two servers, one with your database/active directory/SQL server/Amazing Charts and a second one that handles the thin clients (terminal server).


Personally, I'd keep a separate server/main computer for AC and have a desktop in my office. Many physicians do more than just basic office tasks on their personal computers. Web browsing, watching videos, copying things off cd/dvds, personal printers, digital voice recorders, using viewers for certain test results. So I think it's pretty obvious my recommendation would be a thick client. If you're just starting out, consider getting a server with SBS Essentials and a desktop.


If you knew how old I was, it'd be more the other way around.

thank you "sonny"-- from granpa seriously that really clarifies things for me.

You're welcome ha

That was a great and concise tutorial. To add to it since it is set up so well, since thin clients can access apps on a server via a second server known as a terminal server, then you can see the real usefulness in a terminal server.

A remote office can use computers (hopefully thick ones or regular desktops) and RDC (remote desktop connection) formerly known as RDP (remote desktop protocol) can remote into the terminal server and use AC or any app remotely. Of course, each computer needs a separate CAL.

I've been using a thin client in my office for the past year. I don't miss the big box or it's noise and extra maintenance. Wireless networked printer makes it work fine, as I really hate how printers don't really work off the thin client USB ports.

If you're using an SSD, there's essentially no maintenance. Never have to defrag. Even with a thin client, you have to defrag the server. I just know I wouldn't be able to deal with the limitations of a thin client.

I agree, but to each his/her own. Sandeep, would you need Trim in the situation above?

If your computers set up with the right fans, etc., you won't even know they are on. All Lenovos are like that. Thought the power button was broken the first time I turned it on.

SSD will do garbage collection when idle even if they don't have TRIM. TRIM is just active garbage collection. Windows 7 already has it built in and so do the server operating systems. It's strongly recommended to not defrag SSDs.

Also about the fans, there's ones that you can barely hear. Out of the big 3 HP, Dell, and Lenovo. Lenovo tends to have the highest quality and that's reflected in their price tag ha. But if you want a quiet computer, you can swap out the fans for Scythe Slipstream fans or SilenX fans. You can barely hear them. There's also passively cooled heatsinks. Usually very large, but no noise.

I just did a crazy build where the computer was cooled using full oil submersion in a fish tank which was in another fish tank filled with fish. (I know, I get the craziest requests sometimes ha.) I also put 4 meters of red and blue LEDs in the fish tank. The guy was just speechless when he saw it in action. He thought it was some kind of prank. That computer was dead silent as well.

That was sounds fishy to me. Unless the first fish tank had no water.

The parts in the first tank were submerged in 5 gallons of mineral oil, hence the full oil submersion. Oil isn't conductive so it works. Then I put the first tank in a second larger tank with water and fish. The mineral oil is transparent so it looks like the fish are swimming in the computer.

Sandeep,

Maybe you can't show a picture of that set up but I'd like to see that one, and i never really thought about mineral oil being non-conductive or conductive--mineral oil in the ear to keep wax from getting too hard shouldn't short out the hearing aids, you and Bert keep a guy from getting intellectually stale.

Thanks jimmie,

This is why I live in the easternmost city in the U.S. (Well, over 20,000 residents). Eastport, ME is furthest east

Sandeep has to be pretty close to the westernmost.

But, Chris and Allen were rather intellectual as well today. Good to hear their thoughts.

I got to go look for their thoughts--sorry Chris and Allen!!!
I'm close to the northernmost (continental)--not too far from the land of the maple leaf. Although Brewer is a bit far north too.

I have been to Bellingham quite a lot.

I had to look up where Bellingham is--and did you mean Bellingham Washington?--one of my to do lists is to visit Victoria B.C. I believe right across from Bellingham--but I do really appreciate everyone's insight on the thin/ thick client discussion. It has been quite helpful as I am in the process of getting a server.

Yes, Washington. I am not a big proponent of thin clients, but that is probably more due to my lack of knowledge than anything else. If you aren't familiar with thin clients and you are stepping into the realm of a server, maybe tackling both would be a lot.

Of course, this will bring up a firestorm of thin client backers stating how easy it is to use them. Which would be great as they could help you.

Bert,
I am having my local IT guy look at Sandeep's server recommendation, and looking at just getting a server with a mirrored drive and still use PC's/laptops.

Is he going to build it or order it?

I think leaning towards ordering but I need to get together with him--I at least directed him to your thread about servers to make sure that which ever direction we decide that we model it after your rec's.

Great idea. It'll save you a good chunk and if your IT guy has any problems, just send me a message. Be sure to forward him the videos with the SBS-E instructions as well. Glad to hear the effort put into the recommendations wasn't completely futile haha.

Enjoy installing the recommended Adaptec 6405 drivers. Took me two days!

Sandeep,
I appreciate your help!!!

To clarify, it took me a day and a half.

The OS couldn't find the drivers on the Adaptec CD. So I downloaded the correct drivers from Adaptec, and burned them to a CD. The trick in getting them installed was to insert the ORIGINAL Adaptec CD when Windows SBSE asked for the drivers, THEN insert the new driver CD (that I burned).

Or something like that. It was a terrible experience. I also got an ASUS RT-N66U wireless router that had crappy firmware. Wouldn't save my port-forwarding. Fortunately, the latest driver (3.0.0.3.112) solved that.

Yea it is a bit confusing. You install the driver when you first install at boot time. There's usually a utility on the CD (at least for 3ware), some assume you can figure it out. They want you to copy the files to a floppy (old school) because these have been around for a while and that was the RAID drivers were loaded in the pre-Windows Vista era. Now you can just copy them to a flash drive. Almost every RAID Card manufacturer does the same. It's just assumed to be common knowledge if you're buying on of these cards. It got me for a while as well because I also skipped over the manual like always ha. I actually looked on the CD and saw that it was zipped. Which is why it wasn't working. I believe the manual is incorrect when it says you can use the 3ware CD directly.

The drivers are usually located in a zip file or on a folder on the CD which you extract to a floppy. But nowadays we have flash drives and what not so you can just copy it there. When it asks you for the location of the driver, point to the flash drive.

I didn't read over Adaptec's User Guide, but 3Ware tells you have to make the driver disk first.

I was referring to the Adaptec 6405 RAID controller that you recommended for higher-end builds. See the "Note" in green at the end. Took me awhile to get the driver to load (had to switch between the supplied install disk and the new driver disk I burned).

It's not a CD/DVD/Flash Drive issue. At least not for the 6405. I could browse to the correct directory on the supplied installation CD, but the drivers weren't current. And when I tried to load the NEW drivers that I had downloaded from Adaptec, they weren't recognized.

The only way I finally made it work was to keep the old install CD in the drive till it asked me for the drivers, then replace it with the new driver CD that I had burned, and THEN put the old install CD back in the drive.

Yeah, it was weird...and I almost returned the controller.

I just skipped over the supplied CD and went straight to the driver online and browsed to the driver directly. Then I got an option to pick which model I had. This was the 3Ware 9650SE-4LPML as the supplied CD wasn't working for me either.
The 6405 is better with the 512 MB of cache memory and SATA 6 GBPS ports.

Whatever. We're talking about different RAID controllers and drivers, kiddo.

Sandeep is a great resource, but beware of him ;-)

It's the same principle though regardless ha. I think I also had to do that CD switch. I can't recall though.