Chapter 3: Matters of Protocol

In this chapter, you examine the communications and messages that move around on networks. We tell you more about what senders send and receivers receive while you investigate the sets of rules called protocols that govern how computers exchange information across a network.

In essence, networked communications rely on a shared set of rules for exchanging information and for defining how data looks at the most basic level, such as how to present data digitally (or "What's a one, and what's a zero?"). These rules also dictate the formats for and meanings behind the addresses that indicate where "here" and "there" are on a network, that identify message types and content, and that offer lots of other critical information to boot.

When Computers Communicate

Many of the ways that computers communicate and that humans communicate share common elements. Take a phone call, for example:

• Phone calls use highly formulaic introductions to connect the right speakers on each end of the connection. ("Is this the Phlogiston residence? May I speak to Phil, please ?") Computers take a similar tack for network communications in that a sender often begins by asking the receiver if a conversation can begin, and only after permission is granted does any actual exchange of data occur.

• Taking turns talking on the phone requires careful listening skills and sensitivity to open spaces in the other party's talk, so each party can speak when the opportunity arises. Computers have no intuition, so they exchange explicit signals when one party wants to switch from listening to talking. In fact, some communications techniques allow both parties to talk and listen at the same time!

• Ending a phone conversation can be a matter of mutual agreement or it can involve well-known signals that one party wants to end the conversation. ("I have to let you go no" is a famous human example.) Computers also exchange signals to indicate that a network conversation is ready to end and then conclude by breaking their connection to each other.

• Human possess coping skills to help them recognize unplanned endings to conversations, such as a failed cordless phone battery, driving beyond a cell boundary, or an outright connection failure. They also have the smarts to try again or give up, depending on whether they've satisfied their communication goals. Computers are more simple-minded; they wait until communications resume or a fixed interval of time (called a timeout period ) elapses before recognizing that a connection is dead and that the conversation is over. Then it's up to the application that initiated the link to decide whether to try again or give up.

Understanding the differences between human communications and computer communications can help you understand networking better. The biggest difference, it seems, is that humans can navigate by the seat of their pants far better than computers can.

The secret's in the interpretation

When humans communicate on the phone, what we say (or hear) is always interpreted and often misunderstood. What you think you said isn't always what another person thinks he or she heard you say. Human communication relies on shared rules and meanings as well as a common frame of reference. Computers rely on these same elements to communicate; but because computers can't make judgment calls or use their intuition, these elements must be spelled out completely. Computers can do only what they're programmed to do.

For computers to exchange data, every element must be explicity supplied. Computers can't pick up implications and hidden meanings. To communicate, computers have to begin with complete agreement about the following issues (as stated from a computer's point of view):

• What's my address? How do I learn my address? How do I learn other computers' addresses?

• How do I signal another computer to indicate that I'm ready to send (or receive) a message, that I'm busy, or that I can wait if it's busy?

If you think about the phone system, these issues are the same for humans dialing a telephone and computers dialing a modem. In fact, these questions can be restated as follows :

• What's my phone number? How do I learn my phone number? How do I learn the phone numbers for other parties that I want to reach?

• How do I place a call? How do I recognize a busy signal? How do I get the phone to keep dialing if the number I want to reach is busy? (Note also that the phone system handles busy and ring signals, so both computers and humans can tell when a call is going through and when the party they're trying to reach is busy.)

Agreeing on a set of rules

Building a complete and consistent set of rules for computer communications is a timeconsuming, nitpicky business that's entirely capable of driving most ordinary people bonkers. In the early days of the computer industry, individual companies or groups would put hordes of programmers to work building computer communications programs to solve specific, isolated problems.

But as time went on, programmers realized that this approach produced lots of unique ways for computers to communicate that worked only in the confines of small, isolated technical communities. After the need to communicate spreads farther, serious incompatibilities prevented such communities from exchanging data unless one community willingly gave up its way of communicating and adopted another's way of communicating.

The U.S. government played a key role in bringing order to this network chaos. When the government tried to get computers from Company A to work with computers from Company Z, it soon realized that it had a monster compatibility problem. A consensus soon emerged that a common set of rules for networking would make communication easier. Likewise, early network pioneers quickly learned that networking was difficult, if not downright impossible , when all players didn't follow the same set of rules.

If this tale had a storybook ending, it would be "Today, there's only one set of networking rules that everyone uses wisely and well." Alas, that's not the case. The degree of networking chaos has decreased significantly, but many sets of mutually incompatible networking protocols are still in use because hardware and software vendors try to stay on the "bleeding edge" by inventing new rules as they boldly go where no network has gone before.

 

These fundamental questions must be answered , and they represent just the beginning of a large and complex collection of details that have to be nailed down, codified, and implemented for computers to be able to communicate across a network. The answers to this entire collection of questions are the basis for a set of rules for computer communications; in fact, these rules represent the rules of the road or protocols for networking.