Local Area Network (LAN)|
A LAN is really nothing more
than a wire. When you boil it down to its simplicity, it is a wire that connects
multiple computers into a system to share resources. You have to start by thinking
with "resources" rather than "hardware" because a network changes
all the rules. Until now, you've always thought that C: was your hard disk drive
and D: was your CD drive and E: was your Zip disk drive, your modem is on Com1,
your printer is on LPT1, etc. That is all "hardware" and it is all glued
together by your computer's Operating System right?
The above is a first example
of the Smoke & Mirrors that we're going to be talking about extensively for
the next couple of pages. When it comes to Networking, there is a great yawning
chasm between Reality and Truth. If you run FDISK.EXE on the above hard drives,
you'll see the Truth about all those drive letters, even though you KNOW that they
are C:, D:, E:, F: & G: when you're doing your work because that is Reality.
So, for the rest of this paper, remember that REALITY is what everybody
knows, and TRUTH is something completely else.
The truth is that the above
partitions aren't drives, they're volumes into which you can store your data. It's
important that you understand these new words and their underlying concepts, because
without them you're going to sink into the whirlpool and be totally at the mercy
of your IT (Information Technology) Mgr or your Networking Consultant which is a
dangerous place to be. ;-)
Remember we talked about Resources? Everything we've mentioned so far is a resource that your computer uses to help you do your job. Now we're going to introduce another resource which will change all the rules, the NIC (Network Interface Card) plugs into a slot inside your computer and talks to the wire attached to its outer jack. Whether the wire is 10base2, 10baseT, 10base100, AUI, or FibreOptic doesn't matter to this discussion, we can go into those subjects later if you'd like, suffice it to say that any of these wires supply Ethernet to your computer through the NIC. And that opens the door to a whole array of additional resources that you can exploit to do your job.
You've still got your C: drive on your local machine, and your trusty old Laser printer is still on LPT1, but now you're D:, E:, F:, G: drives are network resources AND you've got a Plotter on LPT2, and a color printer on LPT3 as well as a "Fax machine" on LPT4 which are in truth, totally separate computers down the hall and around the corner that are also being used by your 18 coworkers in the company at the same time that you're using them.
The reality/truth would probably look something like this:
Don't get too lost here! Remember,
none of this truth matters, because the reality is that you open a file on M: drive,
edit it in WordPerfect and print it to the color printer. Your NOS (Network Operating
System) takes care of keeping track of what lies it's been telling you. You just
have to do your job in blissful ignorance of the truth. Kind of like in real life
Upon graduating to more than
three computers (aka workstations or clients) on your network, it is time to consider
a Dedicated File Server. There's a number of reasons for this, but let's get the
nomenclature out of the road first. A "Server" is merely the machine that
makes a Network Resource available to the network. The color printer on "BigBoy"
in Reception in the above example is a network resource and the fact that "BigBoy"
has it attached and makes it available to the network as a resource defines "BigBoy"
as a "Server" because it serves the network. Because
"BigBoy" is also the receptionist's workstation, we define this network
as "Peer-to-Peer" rather than "Client/Server" which we define
as all network resources centralized in a Server that does nothing else but be a
Server, in other words, no one is using it as a workstation. This is admittedly
an oversimplified explanation of the differences between Peer-to-Peer and Client/Server
topology, but it should suffice for the level of work we do most often. If you want
to talk about "Application Servers" and "Directory Services",
you'll have to schedule an individual consultation with one of us, so let's leave
these definitions as is for the rest of this discussion.
Our primary argument for a Dedicated Server isn't performance nearly as much as it is simplicity. It's a LOT easier to administer a network that is telling exactly the same lies to everyone on the network than it is to have to remember which lies receptionist was told vs which lies we're telling the AutoCAD drafters, vs yet another set of lies to Administration and Accounting. With all resources centered in a single server or bank of servers, it is then possible to map out a standard topology for everyone and debugging future problems becomes, if not easy, at least possible.
Performance is another important factor in this type of approach. When the Server's system can be dedicated to just that task, all the computer's horsepower can be assigned to being the server rather than keeping some in reserve for the operator to use as a workstation and the entire network feels that advantage. However that advantage pales into insignificance when we get to the redundancy factor. All magnetic media is by definition ephemeral. It goes away with extremely little warning and sometimes none at all. Therefore, a system of backing up your data on a regular basis is vital. With a central Server containing all the important data of your office, it is possible to do regular backups automagically and when you graduate to 10 or more workstations we'll start hammering on you to add a "mirrored" Backserver into the mix for yet another layer of redundancy. Backups are only as good as they get done, so we strive to make backups as uninvasive to your operation as possible, thereby ensuring that they are available when you need them. PMAco975seSandyBv