3Com Dual 56K manual Becomes, Is written as

140APPENDIX A: NETWORKING PRIMER

What is an IP Address? An IP address is a numeric value of a set form—32 bits, or binary digits of either one or zero—which is used within a TCP/IP network to uniquely identify the location of every device within the network for purposes of routing data. Each bit is in actuality a digital signal representing either a one or zero.

Because of the lengthy form of binary values when written in their binary form, the convention followed when writing IP addresses is to divide the 32-bit address string into four sets of eight bits, or four bytes (one byte is generally made up of eight bits). Each set of eight bits is then separated by a period, and each resulting binary value is then converted into its decimal equivalent. In this way, for example, the IP address 11000000101010000000000100000001

. . . becomes 11000000.10101000.00000001.00000001

. . . and is written as 192.168.1.1

Much as a postal address uses several levels of identification to pinpoint an exact location (country, zip code, state, street), an IP address incorporates two levels of identification within its 32-bit structure. There is a network number or address, which identifies the broader level of identification (much like a zip code region), and a host number or address, which pinpoints the exact location of a device within the broader network. Specifically, within each set of 32 bits, the first segment of the bits serves to identify the network on which a device resides, and the remaining bits identify the location of that device (or host) on that network.

The division between network and host is significant because public (i.e., Internet) routers only route to the level of the network address. In other words, “network” in this context has a very specific and circumscribed meaning. Beyond this network level, it is up to a private organization’s internal routers to pinpoint a host’s exact location. (This has significance for the use of subnet masks, which are explained below.) All this might be analogous to the idea that, in a first round of mail sorting, for efficiency’s sake, mail is sorted only to the level of the zip code, whereupon a second group of mail sorters starts the process over again at the next level down.

Although it might be logical to assume that the first sixteen bits of an IP address identifies the network and the last sixteen the host, this is not necessarily the case. Rather, although the breakdown always follows the pattern, first part = network; second part = host, the number of bits used to identify “network vs. host” varies, depending upon the Class type into which the network falls. A network’s Class type—Class A, B, or C (Class D and E also exist for special purposes)—is determined by its size. The establishment of these different classes of networks has to do with the early history of Internet design planning, but the end result is that, from largest to smallest, network types are referred to as Class A, Class B, and Class C.

Within each class, then, a set number of bits identifies the network address, and a set number identifies the host address. In a Class C network, for example, the first 24 bits identify the network (in actuality, the first three bits are always set to 1-1-0, and the next 21 identify the network). The final eight bits identify the host.

Because these final eight bits can yield a possible combination of 256 different binary numbers (28 = 256), the number of host devices that can reside on any one Class C network is limited to 256 (actually, 254, because the all-0’s and all-1’s are