Not sure what info was available from the link given, or if it has the same info as I am including here, so I'll toss in some quick-ees:
LAN / WAN Tech 101
... Please take notes as there will be a quiz at the end of the discussion!
Term - Server [simplified / general meaning]:
Any Networked Computer that provides some type of service to other Computers on the Network.
Types of Servers:
[*]Disk Servers: Primarily a file storage device. File Serving first began with the Disk Server approach. This can also cover Networks using Dumb Terminal type nodes.
Downside of this type of Server is that each node (workstation on the network) maps the FAT of the server's fixed disk drive (Hard Drive) at the time of Boot-up, so when the FAT gets changed by normal file use, no one can tell what the disk drive looks like without a reboot.
Instead of sharing the "Physical" disk to the Network, a file server shares the Storage Area. It works kind of like disk serving, but instead of each client seeing the server's entire drive (and FAT), the drive is partitioned and each client maps a drive letter (or Unicode path) on the shared volume, which appears like another fixed disk - local to the client machine. This "fixes" a FAT to the drive letter, so each partition on the server's drive is like a separate drive on the client machine(s). When it (the FAT) changes, it's due to the file work done by that particular client - hence it's automatically kept current.
Print Servers are either a regular, run-of-the-mill type Computer with (some) CPU power, or simple stand-alone print server devices. For the "PC" type server, a nice 80486 DX2 will work fine for Print Serving.
This concept allows multiple clients to print to one shared printer. This also can allow multiple printers to be shared by clients.
Print jobs are "Spooled" to the print server, which frees up the client's machine from printing tasks.
Simply, these servers are MODEM Pooling and Proxy servers. All clients share MODEMS connected to the Comm Server for Communications type service, such as Faxing (In and Out), WAN applications, Internet work and Internet + WAN type E-Mail.
Can be used with common Analog services (MODEMs), or use Leased Lines (such as ISDN, T1,2,3 and such), or with a certain flavor of DSL (ADSL, SDSL, HDSL, etc.).
As the name applies, instead of each client holding the working copy of the Application, the Application Server holds the Application and each client gets a "Copy" of the App' when the client "requests" one. (Applications are the programs you use to do Corn-Puter / Nerd work, such as Excel, Lotus 1-2-3, Word, Adobe Acrobat, Netscape, AutoCAD, etc.)
This keeps the App' from becoming corrupted, plus helps keep the most recent and less buggy version for use.
One of the simpler reasons for this has to do with Security (mostly in the form of theft and piracy protection).
This method requires the Company's Net Geek (Network Administrator) to setup clients with "Instances" and all that junk for all LAN work. This also requires a net geek to be in the "Nerd Room" (Administration Room) at all times to monitor network usage (I am very green with App' server tech, so correct me if I am wrong with the monitor protocol!)
Peer-To-Peer networks may or maynot have a dedicated file server machine, as each machine is just the same as any other one. Client Server networking does not automatically dictate one certain machine as a "Dedicated Server" - all machines are treated equal. Setting up instances and stuff like that just limits the ability for clients to access shared resources.
Hope I have given you some helpful basics as to "Whaddaheck Is A Server, Anyway???" I have covered this a lot, and it gets a little easier to explain each time, so if the above info does not come across very clear, just wait for me to discuss it a few more times and the stuff should clear up a lot more