Introduction to Networks

Contents:

MCSE Links
Links to MCSE sites with
reviews and summary of
contents.

Exam Notes
Exam notes for subjects

Win NT Resources
Extensive NT resources,
online mags, product reviews...

Technical Articles
NT articles, incl DNS, DHCP,
WINS, TCPIP etc.

Downloads
For sample exams for both MCSE and MCSD papers

About me
Unimportant things about the writer

Nowadays you hear a lot about networks. What is it?? It would probably bring to mind a web-like system of interconnected ‘things’. On a broad term, those ‘things’ can be anything, from computers, to telephones to people. As long as they are connected in a meaningful way in order to exchange data, they are on a network.

Why Network?

Why would people want to connect their machines together? One of the possible reasons is that people have this innate greed to share in other peoples possession, and so came about the ‘sharing’ concept. You share my hard disk space, I share your printer.

It’s a more efficient concept as well. A group of people could share a printer to print out documents, and that would save on cost as well. Not only that, but networking would mean enabling common administration and security, especially if you are on a client-server model. You could get by with having loads of stand-alone machines (computers that are isolated from each other, i.e. not connected in any physical way), but that would be pointless really.

Types of Networks

There are three kinds of network models. All of which will be discussed below :

Peer-to-Peer Computing Model

First of all, we have the peer-to-peer kind. Another name for this would be Distributed Computing Model. This would be analogous to the classroom environment, but without an instructor. Each student shares gossip and/or information about who did what where and when. No one is ‘higher’ or ‘more superior’ than the other in terms of hierarchy or knowledge-wise. In a peer-to-peer network, each computer may share its own data with someone else subject to them knowing the password (obviously for security reasons you don’t want just anybody to read your hard disk).

This is also sometimes called a workgroup, and is ideal for small companies with fewer than ten or fifteen users where security is not a critical issue. Shared resources like printers may be deployed in such a network, where the software to achieve networkability is not very complex. A machine with a Windows for Workgroup, Windows 95, or Windows NT operating system would be able to be networked easily as they have built-in peer to peer functionality.

Other operating systems that cater to this kind of model includes Novell Personal NetWare, AppleTalk (the networking system for Apple Macintosh computers), and Artisoft LANtastic.

And a peer-to-peer computing model involves the following :

• Multiple computers capable of operating independently
• Tasks can be completed locally on various computers
• Networks enable the computers to exchange data and services but do not provide processing assistance.

Client-Server Computing Model

Imagine each person connected to each other by imaginary channels of communication (i.e. talking) and there you have a network of communication. If the students are requesting and receiving instruction and knowledge from the instructor, then they can be considered as ‘clients’ and the instructor as ‘server’. In the computing world, that is exactly what it is. A client computer in a network basically just requests for services or data from a server.

That is a network in a client-server environment. It is the accepted norm in most companies now, to have their network in a client-server way. It is also called Collaborative or Cooperative Computing Model.

There are basically three kinds of client-server deployment. One of them is file server model and the other is application server model and lastly we have the printer server. The file-server just does that, stores files for the use of clients accessing them from a network location. This maintains a central location for files and documents, which makes it easier for the administrator to do back ups.

The Application server does part of the processing for the client machines, whereas the file server just stores the data (in a form of a file) for the client. The client might query the server, for example, "Please list all users in the customer database whose name begins with ‘Smith’". The server would do the searching and present the final answer to the client.

A print server on the other hand, just manages access to a shared printer, making it accessible to users at other networked machines.

In short, the client-server model has the following features :

• Multiple computers cooperating to perform a task
• A network that enables the computers to exchange data and services
• Software designed to take advantage of the collaborative environment
• Server can provide higher security and administration purposes when used to maintain users/security.

Host-Based Computing Model

During the old days of yore, when computers first came out, they were huge clunky things, large room-sized humongous Big Irons servicing people sitting at dumb terminals. What is the difference between this host-based system and the client-server environment? Well, in a host based system, the dumb terminals are just that, dumb. They don’t think. They listen and they do, something like a robot really, without thinking. The host does all the thinking for them. This system is called the Centralized Computing Model. Everything is centralized. Whereas in a client-server environment, the clients can do some processing on their own as well, without taxing the server.

Basically in summary, a host-based system has the following characteristics :

• All the processing takes place in the central, mainframe computer (also called the host)
• Terminals (dumb ones, not very smart ones) are connected to the central computer and function only as input/output devices
• Networks may be employed to interconnect two or more mainframe computers. Terminals connect only to the mainframe, never to each other.

Let me give you an example of how a client-server environment might work in a classroom.

Student (Client) : Teacher, how should I do this sum?
(Request info)

Teacher (Server) : Well you could try this method with this information.
(Service request)

Student (Client) : Oh ok, I am trying out this now, and I’ve got the answer!
(Does own processing and finishes computing)

Here we have a host-based classroom environment:

Student (Dumb terminal) : Teacher, what is 2+2?
(Request info)

Teacher (Host) : 4, you dumb thing.
(Service request)

Student (Dumb Terminal) : Oh.
(Ding! Lightbulb)

Notice that the host-based system is similar to the application server based model of the client-server system. There isn’t a lot of difference there really, except in the client-server model the client has the processing power to do other tasks by itself, than just request processed data from the server.

There you go. All clear now? That’s the difference between client-server based and host-based systems. Why is the trend moving towards client-server system now? Well, other than removing workload off the host to do other stuff, the ‘smarter’ client (which has its own CPU, i.e. brain as compared to a dumb terminal who has none) can do it’s own thing too.  

Networks come in all shapes and sizes. Therefore, it is easier to classify them into geographical size, as those networks of similar sizes share characteristics. The two types of networks are

• Local Area Networks (LANs), and
• Wide Area Networks (WANs).

Local Area Networks (LANs)

When a group of computer that is interconnected in such a way that they are geographically close to each other, like in the same building or campus, it is classified as a LAN. The nodes in these type of network normally use only one type of transmission medium – cabling.

LANs are characterized by the following features :

• They transfer data at high speed
• They exist in a limited geographical area
• Their technology is generally less expensive

Wide Area Networks (WANs)

When you have a need to connect two LANs together, for example if you have two branches of the same organization at different sites, having a geographical distance between them to be connected, you have a wide area network (WAN). A WAN may be located in a state or country, or even be interconnected around the world.

A WAN has the following features :

• They can span a distance of unlimited geographical area
• Due to the distance the data has to travel from one site to another, they are then more susceptible to errors
• They interconnect multiple LANs
• They are more sophisticated and complex than LANs.
• Their technology is expensive.

If you have any comments, please feel free to email me at foogy@geocities.com
Copyrighted 1997 Lim Pei Mun J
Last updated : 30 April 1998

Thank you (kissy kissy *smack* *smootch*) for hosting my page
You can also get your own Free Home Page