The need to share information and resources among different computers has led to linked computer systems, called networks, in which computers are connected so that data can be transferred from machine to machine. In these networks, computer users can exchange messages and share resources—such as printing capabilities, software packages, and data storage facilities—that are scattered throughout the system. The underlying software required to support such applications has grown from simple utility packages into an expanding system of network software that provides a sophisticated network-wide infrastructure. In a sense, network software is evolving into a network-wide operating system. In this chapter we will explore this expanding field of computer science.
Network Fundamentals
We begin our study of networks by introducing a variety of basic networking concepts.
Network Classifications
A computer network is often classified as being either a local area network (LAN), a metropolitan area network (MAN), or a wide area network (WAN). A LAN normally consists of a collection of computers in a single building or building complex. For example, the computers on a university campus or those in a manufacturing plant might be connected by a LAN. A MAN is a network of intermediate size, such as one spanning a local community. A WAN links machines over a greater distance—perhaps in neighboring cities or on opposite sides of the world.
Another means of classifying networks is based on whether the network’s internal operation is based on designs that are in the public domain or on innovations owned and controlled by a particular entity such as an individual or a corporation. A network of the former type is called an open network; a network of the latter type is called a closed, or sometimes a proprietary, network. Open network designs are freely circulated and often grow in popularity to the point that they ultimately prevail over proprietary approaches whose applications are restricted by license fees and contract conditions.
The Internet is an open system. In particular, communication throughout the Internet is governed by an open collection of standards known as the TCP/IP protocol suite, which is the subject of Section 4.4. Anyone is free to use these standards without paying fees or signing license agreements. In contrast, a company such as Novell Inc. might develop proprietary systems for which it chooses to maintain ownership rights, allowing the company to draw income from selling or leasing these products. Still another way of classifying networks is based on the topology of the network, which refers to the pattern in which the machines are connected. Two of the more popular topologies are the bus, in which the machines are all connected to a common communication line called a bus, and the star, in which one machine serves as a central focal point to which all the others are connected. The bus topology was popularized in the 1990s when it was implemented under a set of standards known as Ethernet, and Ethernet networks remain one of the most popular networking systems in use today.
The star topology has roots as far back as the 1970s. It evolved from the paradigm of a large central computer serving many users. As the simple terminals employed by these users grew into small computers themselves, a star network emerged. Today, the star configuration is popular in wireless networks where communication is conducted by means of radio broadcast and the central machine, called the access point (AP), serves as a focal point around which all communication is coordinated.
The difference between a bus network and a star network is not always, obvious by the physical arrangement of equipment. The distinction is whether the machines in the network envision themselves as communicating directly with each other over a common bus or indirectly through an intermediary central machine. For instance, a bus network might not appear as a long bus
from which computers are connected over short links. Instead, it may have a very short bus with long links to the individual machines, meaning that the network would look more like a star. Indeed,
sometimes a bus network is created by running links from each computer to a central location where they are connected to a device called a hub. This hub is little more than a very short bus. All it does is relay any signal it receives (with perhaps some amplification) back out to all the machines connected to it. The result is a network that looks like a star network although it operates like a busnetwork.
Protocols
For a network to function reliably, it is important to establish rules by which activities are conducted. Such rules are called protocols. By developing and adopting protocol standards, vendors are able to build products for network applications that are compatible with products from other vendors. Thus, the development of protocol standards is an indispensable process in the development of networking
technologies. As an introduction to the protocol concept, let us consider the problem of coordinating the transmission of messages among computers in a network. Without rules governing this communication, all the computers might insist on transmitting messages at the same time or fail to assist other machines when that assistance is required.
In a bus network based on the Ethernet standards, the right to transmit messages is controlled by the protocol known as Carrier Sense, Multiple Access with Collision Detection (CSMA/CD). This protocol dictates that each message be broadcast to all the machines on the bus (Figure 4.2). Each machine monitors all the messages but keeps only those addressed to itself. To transmit a message, a machine waits until the bus is silent, and at this time it begins transmitting while continuing to monitor the bus. If another machine also begins transmitting, both machines detect the clash and pause for a brief, independently random period of time before trying to transmit again. The result is a system similar to that used by a small group of people in a conversation. If two people start to talk at once, they both stop. The difference is that people might go through a series such as, “I’m sorry, what were you going to say?”, “No, no. You go first,” whereas under the CSMA/CD protocol each machine merely tries again later.
Note that CSMA/CD is not compatible with wireless star networks in which all machines communicate through a central AP. This is because a machine may be unable to detect that its transmissions are colliding with those of another. For example, the machine may not hear the other because its own signal drowns out that of the other machine. Another cause might be that the signals from the different machines are blocked from each other by objects or distance even though
they can all communicate with the central AP (a condition known as the hidden terminal problem. The result is that wireless networks adopt the policy of trying to avoid collisions rather than trying to detect them. Such policies are classified as Carrier Sense, Multiple Access with Collision Avoidance
(CSMA/CA), many of which are standardized by IEEE within the protocols defined in IEEE 802.11 and commonly referred to as WiFi. We emphasize that collision avoidance protocols are designed to avoid collisions and may not eliminate them completely. When collisions do occur, messages must be retransmitted.
The most common approach to collision avoidance is based on giving advantage to machines that have already been waiting for an opportunity to transmit. The protocol used is similar to Ethernet’s CSMA/CD. The basic difference is that when a machine first needs to transmit a message and finds the communication channel silent, it does not start transmitting immediately. Instead, it waits for a
short period of time and then starts transmitting only if the channel has remained silent throughout that period. If a busy channel is experienced during this process, the machine waits for a randomly determined period before trying again. Once this period is exhausted, the machine is allowed to claim a silent channel without hesitation. This means that collisions between “newcomers” and those that have already been waiting are avoided because a “newcomer” is not allowed to claim a silent channel until any machine that has been waiting is given the opportunity to start.
Sharing is Sexy !
About the Author
0 comments: