|
|
| Previous | Table of Contents | Next |
TCP/IP is a set of data communications protocols. These protocols allow for the routing of information from one machine to another, the delivery of e-mail and news, even the use of remote login capabilities.The name TCP/IP refers to the two major protocols, Transmission Control Protocol and Internet Protocol. Although there are many other protocols that provide services that operate over TCP/IP, these are the most common.
Internetworking with TCP/IP has been around for many years—almost as many years as Unix has been available. TCP/IP, or Transmission Control Protocol/Internet Protocol, grew out of the work that was done with the Defense Advanced Research Projects Agency, or DARPA. In 1969, DARPA sponsored a project that became known as the ARPANET. This network mainly provided high-bandwidth connectivity between the major computing sites in government, educational, and research laboratories.
The ARPANET provided those users with the ability to transfer e-mail and files from one site to another, while DARPA provided the research funding for the entire project. Through the evolution of the project, it became clear that a wide range of benefits and advantages were available, and that it was possible to provide cross-country network links.
During the 1970s, DARPA continued to fund and encourage research on the ARPANET, which consisted chiefly of point-to-point leased line interconnections. DARPA also started pushing for research into alternate forms of communication links, such as satellites and radio. It was during this time that the framework for a common set of networking technologies started to form. The result was TCP/IP. In an attempt to increase acceptance and use of these protocols, DARPA provided a low-cost implementation of them to the user community. This implementation was targeted chiefly at the University of California at Berkeley’s BSD Unix implementation.
DARPA funded the creation of the company Bolt Beranek and Newman Inc. (BBN) to develop the implementation of TCP/IP on BSD Unix. This development project came at the time when many sites were in the process of adopting and developing local area network technologies, which were based closely on extensions of the previous single computer environments that were already in use. By January 1983, all the computers connected to the ARPANET were running the new TCP/IP protocols. In addition, many sites that were not connected to the ARPANET also were using the TCP/IP protocols.
Because the ARPANET generally was limited to a select group of government departments and agencies, the National Science Foundation created the NSFNet that also was using the successful ARPANET protocols. This network, which in some ways was an extension of the ARPANET, consisted of a backbone network connecting all the super-computer centers within the United States and a series of smaller networks that were then connected to the NSFNet backbone.
Because of the approaches taken with NSFNet, numerous network topologies are available, and TCP/IP is not restricted to any single one. This means that TCP/IP can run on token ring, Ethernet, various bus topologies, point-to-point leased lines, and more. However, TCP/IP has been closely linked with Ethernet—so much so that the two were used almost interchangeably.
Since that time, the use of TCP/IP has increased at a phenomenal rate, and the number of connections to the Internet, or this global network of networks, has also increased at an almost exponential rate. Countless numbers of people are making a living off the Internet, and with the current trends in information dissemination, it likely will touch the lives of every person in the developed world at some time.
TCP/IP, however, is not a single protocol. In fact, it consists of a number of protocols, each providing some very specific services. The remainder of this chapter examines how addressing is performed in TCP/IP, network configuration, the files controlling how TCP/IP can be used, and many of the various administrative commands and daemons.
Note: A daemon is a program that runs to perform a specific function. Unlike many commands that execute and exit, a daemon performs its work and waits for more. For example, sendmail is a daemon. It remains active even if there is no mail to be processed.
Each machine on the Internet must have a distinctly different address, like your postal address, so that information destined for it can be successfully delivered. This address scheme is controlled by the Internet Protocol (IP).
Each machine has its own IP address, and that IP address consists of two parts: the network portion and the host portion. The network part of the address is used to describe the network on which the host resides, and the host portion is used to identify the particular host. To ensure that network addresses are unique, a central agency is responsible for the assignment of those addresses.
Because the original Internet designers did not know how the Internet would grow, they decided to design an address scheme flexible enough to handle a larger network with many hosts or a smaller network with only a few hosts. This addressing scheme introduces address classes, of which there are four.
IP addresses can be expressed in several different forms. First is the dotted decimal notation, which shows a decimal number with each byte separated by a period, as in 192.139.234.102. Alternatively, this address also can be expressed as a single hexadecimal number such as 0xC08BEA66. The most commonly used address format, however, is the dotted decimal notation.
As mentioned, there are four major address classes: class A, B, C, and D. Classes A, B, and C are used to identify the computers that share a common network. A class D, or multicast address, is used to identify a set of computers that all share a common protocol. Because the first three classes are more commonly used, this chapter focuses on them. Regardless of the address class, each address consists of 32 bits, or 4 bytes. Each byte is commonly referred to as an octet, so an IP address consists of four octets.
Each octet can have a value from 0 to 255. Certain values, however, have a special meaning that is shown in table 1.1 later in this chapter.
| Previous | Table of Contents | Next |