IP 101: All About IP Addresses
By Chris Lewis The key to understanding IP, and all of the issues related to IP, is knowing what a routing table looks like and the effects each IP topic has on the entries in a routing table. To begin with, let's review the basics. IP addresses are 32 bit numbers, most commonly represented in dotted decimal notation (xxx.xxx.xxx.xxx). Each decimal number represents eight bits of binary data, and therefore can have a decimal value between 0 and 255. IP addresses most commonly come as class A, B, or C. It's the value of the first number of the IP address that determines the class to which a given IP address belongs. Class D addresses are used for multi-cast applications.
(For a full explanation of class D addresses, refer to "Diving Through the Layers" .) The range of values for these classes are given below.
Class Range Allocation A 1-126 N.H.H.H B 128-191 N.N.H.H C 192-223 N.N.N.H D 224-239 Not applicable
The class of an address defines which portion of the address identifies the Network number and which portion identifies the Host, as illustrated above, as N and H.
So, without any subnetting (which we will come to a little later), a routing table will keep track of a) network numbers, b) the next hop router to use to get to that network, and c) the interface this next hop router is reachable through. A simple network with the corresponding routing table for a Cisco router is illustrated below.
C 18.104.22.168 directly connected Ethernet 0 C 10.0.0.0 directly connected Token-ring 1 C 22.214.171.124 directly connected Ethernet 1 I 126.96.36.199 via 188.8.131.52 Ethernet 1Since Cisco doesn't give headings for these columns, you need to know what each column consists of. The first column of the routing table indicates how the network number was discovered. C stands for Connected and I indicates the network was learned from the IGRP routing protocol. For a full description of the routing table as it appears in a UNIX host and a Cisco router, refer to "Should RIP Rest In Peace" .
The important thing to realize is that while a routing table keeps track of network numbers, no one assigns a network number to any piece of equipment. Every interface of a router or host connected on the network must have an IP address and a subnet mask defined (many pieces of equipment will assign a default subnet mask if none is applied). From this IP address and subnet mask, the network number is derived by the IP stack and tracked in the routing table.
(This is the exact opposite of what happens in a NetWare network. In NetWare, you assign a network number to a server LAN card, which is used by all workstations on that wire. The workstations use MAC addresses as IPX node numbers.)
Routing tables can get very large. Internet backbone routers can have over 40,000 routes defined in them. In most corporate networks, the routing table is much smaller, as there are not so many subnets that need to be reached.
Many large routers, particulary internet routers, use a method called Classless Interdomain Routing (CIDR) to reduce the number of entries a router needs in its routing table. If we imagine, for instance, that all the Class C addresses that start with the value 194 are allocated for use in Europe, it would significantly reduce the number of entries in Internet routers in the US if there was only one entry for all these class C addresses, rather than a separate entry in the routing table for each one. CIDR works if (as in this example) all the networks with the first octet value of 194 are physically located in one area of the network.
IP addresses are used to deliver packets of data across a network and have what is termed end-to-end significance. This means that the source and destination IP address remains constant as the packet traverses a network. Each time a packet travels through a router, the router will reference it's routing table to see if it can match the network number of the destination IP address with an entry in its routing table. If a match is found, the packet is forwarded to the next hop router for the destination network in question (note that a router does not necessarily know the complete path from source to destination--it just knows the next hop router to go to). If a match is not f ound, one of two things happens. The packet may be forwarded to the router defined as the default gateway, or the packet may be dropped by the router. (In the language of TCP/IP, a gateway is a router.)
Packets are forwarded to a default router in the belief that the default router has more network information in its routing table and will therefore be able to route the packet correctly on to its final destination. This is typically used when connecting a LAN with PCs on it to the Internet. Each PC will have the router that connects the LAN to the Internet defined as its default gateway.
A default gateway is seen in a routing table of a host as follows: the default route 0.0.0.0 will be listed as the destination network, and the IP address of the default gateway will be listed as the next hop router.
If the source and destination IP addresses remain constant as the packet works its way through the network, how is the next hop router addressed? In a LAN environment this is handled by the MAC (Media Access Control) address, as illustrated below. The key point is that the MAC addresses will change every time a packet travels though a router, however, the IP addresses will remain constant.
PC1 Router E0 Router E1 PC2 MAC Address M1 M2 M3 M4 Software (IP) address 11 12 13 14 A packet sent from PC1 to PC2 will look like this at point A: Destination Source Destination Source Data MAC MAC IP IP M2 M1 14 11 1001001 A packet sent from PC1 to PC2 will look like this at point B: Destination Source Destination Source Data MAC MAC IP IP M4 M3 14 11 1001001