It is probably not a really correct palce to put up question but possibly it is efficient.I have several question regarding server quality:
1) Why server tends to be in rack form. Is it that they can be stacked up
2) Why Xeon is prefered to Pentium as CPu for server. Some says that Pentium can not run continously for several days but is it a plausible reason.
1) they are put in rack form mostly because of space concerns. it also makes it easyer if there are lots of servers in a rack vs lots of servers piled ontop of eachother and could fall over easily. in a rack station you can take the server from the botton of the stack and remove it from the pile without disturbing the other servers on top.
2) xeon is just better suited for servers. im not 100% but i think the reasons are: supports multiple cpu, can address more ram, more cache, and i think they are 64 bit. they are much more expensive though.
go tell your alien brothers, that ronnie cordova says they're gay!!! <A HREF="http://sockbaby.com" target="_new"> sock baby </A>
Exactly, they use Xeon only because it supports multiple CPU and have more cache. You can have as much memory with Pentium as with Xeon and both are 64 bit now...
The ammount of memory might be another plus with the introduction of FBDimm.
Also server motherboard are specifically made for XEON and in the near future should include Dual Front Side Bus wich should help minimize the Memory bandwith bottleneck!
If reliability is the only concern, should normal mainboard with RAID for data protection do the job. From your answer, it seems server is expensive mostly for performance concern
Server parts are usually of higher quality. Taking Hard drive is the best example! SCSI hard drive are much more reliable. Simply because they're made to last long and to run 24/7 and money is not the main concern. These dont apply to consumer products.
Server are primarly geared toward reliability and to minimize downtime. The performance, generally, is'nt the primary concern(not always true) and money is not really a problem because some downtime will cost you many times more money! Now since server are made to "serve" many user/process/whatever, they mostly have to be efficient in doing parrallel work, thats why you use mutliple CPU and HD's,etc. It gives you reliability and solid parrallel performance!
So that why you see server with dual- everything, from powersupply to cpu, to hd and NIC. All that stuff is hot-swapable so you can replace the defective part without interupting service.
Those are the main reason why you dont use "normal" mainboard because they dont support all this hot-swapping and reductancy.