A.1 An Introduction to Networking Concepts

The general purpose of a network is to transmit information between two or more devices. This usually consists of one system sending a request for information to another system, which then acts upon the request and returns some sort of information back to the requesting system. Sometimes these systems are computers, and sometimes not; they could also be printers, bank teller machines, or telephones. Sometimes these systems are on the same piece of wire, and sometimes they are located on different continents, connected via the Internet or some other global network.

In order to successfully deliver information between the devices on a network, several steps must occur:

1. The originating system has to package the information in a manner which both systems understand.
   
   
2. The sender must then deliver the package to the destination, using techniques that are commonly understood by the systems and the network alike (these packaging and delivery functions are defined as "protocols").
   
   
3. The destination system, upon receiving the package, must check for any errors which may have incurred during transmission.
   
   
4. It must then unpack the package, and pass the data to a local application for processing.
   
   
5. If any information is to be returned to the first system, the above process must be repeated.
   

Although this process is oversimplified somewhat, it describes the basic operation of most communication technologies. Where things start to get complicated is in the differences between the various technologies and products that provide these functions.

Since most network technologies are designed for a specific use, they tend to be highly- optimized for specific environments. This optimization results in specific benefits (and liabilities) that are a direct result of the design goals. For example, modem cables and printer cables are extremely different entities, as are the mechanisms used to provide services across them, although both provide "network" services.

Modems typically use serial cables, which are designed to transmit data one bit a time. While slow, the one-bit-at-a-time design is necessary for devices like modems that rely on the historically noisy public telephone network for data transmission. Every bit needs to be verified for accuracy, so they are sent as single units.

Conversely, printers are typically attached directly to a PC and do not encounter much network interference. This allows for the use of parallel communication cables which are able to transmit multiple bits of information simultaneously. Because they do not need to conduct much error checking, they can transmit much more information simultaneously.

Figure A.1 below illustrates the difference between networks based on serial cables and parallel cables:

Figure A.1 Serial cables send data one bit at a time, while parallel cables send data eight (or more) bits at a time.

As could be expected, parallel cables are much faster than serial cables. However, the applications and services which use serial cables are much more robust than those that use parallel cables, because they have to be able to deal with more signaling noise and errors.

These are fairly big differences, even though the physical cables might look the same to the untrained eye. The systems and applications that use these cables and the related protocols are well aware of the differences however, and are optimized around them.

The more complicated network technologies such as Novell's IPX, IBM's SNA, Microsoft's NetBEUI, Apple's AppleTalk and the Internet's TCP/IP are even more diverse, and much more complicated.

Copyright © 1997, Unoverica Corporation. All rights reserved. Unauthorized use prohibited.