This chapter introduces basic ideas about how telecommunications take place. The objectives important to this chapter are:
Concepts:The basic definitions for terms are given early in the chapter. An architecture is defined on page 77 as a "plan or direction" that specifies "what" we will do, but not how. This means that a telecommunications architecture tells us what a piece of hardware or software must do, but not how to do it.
A standard is defined on page 78 as a set of rules that tell us how to do a given thing.
As systems developed, sold by many different companies, it became confusing and difficult to understand and to get different systems to interact. This created the need for both architectures and standards.
On page 82, the author discusses the V. and X. standards of the ITU-T, an international standards organization. The V. standards have to do with attaching digital equipment to analog communications lines. The X. standards have to do with attaching digital equipment to digital communications lines. If you have a modem in your computer at home, it probably supports a V. standard, if you attach to a standard telephone line. If you are a subscriber to a digital service, you may have a modem that supports X. standards.
(Mr. Rowe discusses analog and digital signals in chapter 5. He gets it wrong, so I will correct it in my notes there.)
On page 83, the discussion is about standards. They are good in that the consumer need only know that a standard is supported by a product in order to decide to buy it. For instance, the television signal standard is different in Japan than it is in the U.S., so a consumer needs to buy a receiver that will support the standard in use at the location where it will be used.
On the other hand, a standard takes a long time to be approved by standards organizations. The digital television standard for the US has been in committee for years and is still unavailable as a product you can purchase.
Two major network communications architectures are presented in this chapter. The first is the IS0-OSI model. Remember, this is an architecture, so it specifies what, not how. It divides the tasks that must be done in network communications into seven levels or layers. Each layer must be implemented by software or hardware that carries out the tasks.
Think of the model as specifying what happens to a message being sent across a network. The originator of the message begins the process of sending it to the receiver. The message is processed by each of the seven layers of this model, starting at Layer 7, and working down to Layer 1, before the message even leaves the originator's computer. The receiver's computer gets the message, and processes it starting at Layer 1 and working up to Layer 7, then can display the message to the receiver.
The seven layers of the ISO-OSI model are:
Note the mention on page 85 that the bottom three layers of this model represent the X. 25 standard for networks. This remind you that we are talking about all digital equipment.
A brief discussion of each of the ISO-OSI layers is found on pages 86 through 88. In my notes at this site for Novell Network Technologies, there is an entire chapter on each layer.
On pages 88 and 89, the author shows us another model, the TCP/IP
model. This is the communications model used to create the Internet.
This model has five layers:
The tasks carried out by these layers are similar to the ones carried out by the ISO-OSI layers, but there is not a neat correspondence.
On page 90, several proprietary standards are presented. These are standards created by single companies for their own use. The idea being that if they could do it better, and theirs did not work with the others, then the world would buy only their products. This is not how the world works, however, so most of these standards are no longer important.
Finally, the author make a few observations about layered architectures. They are meant to be modular, and to work well with replacements. That is, it should continue to work, even if you upgrade some of the software that serves some of the layers. (Like getting a new web browser.) However, this also means that some intelligence is required of the software in these layers, since they have to figure out how to interact with each other, and since almost all the processing takes place at each computer. Back when most users had terminals that connected to mainframes, this was less possible.