multi-seat vs. thin client multi-user systems

Question

I want to understand the theoretical background of multi-user system (how it works). I found we can implement the idea using two methods. 1.Multi-seat 2.Thin client What are the differences between them. And should we install OS in thin client while using in multi-user system. I found thin-clients may not have hard disks. Please help me out..

Answer 1

The basic idea is one of resource utilisation. For example; for a typical office scenario you might buy 10 computers for 10 users at $2000 each, and while they're being used they'll spend most of their time waiting for keyboard or mouse and only actually use about 10% of the CPU, etc. It's wasteful.

For multi-seat; you'd throw some more video cards into the computer and get a USB hub and plug more keyboards in. You might end up with 2 computers at $3000 each with 5 people per computer (or $6000 total cost instead of $20000); and if its done right the users won't really notice the difference. Of course it's a little hard to cram 5 video cards into a single machine (and hard to get 5 users close enough for everything to reach), so this has some practical limitations.

For thin client; you shift the video and keyboard into a little box (the thin client) that communicates with a server over the network. The server runs the applications, etc; and the thin clients don't need to do much processing or anything (and don't need hard drives) and can be really cheap. It will need more networking bandwidth (because the network is now carrying video traffic for all the users); but you can shift the expensive (and often noisy) server into a back room; and because it's centralised it makes maintenance (backups, upgrades, etc) easier. In this case you might spend $4000 on a server and $1000 on 10 thin clients and $1000 on networking (or $6000 total cost instead of the original $20000).