Notes on IP Routing

Since IP is designed as a node-centric networking protocol, every device has equal access to the network. In this model, any device can communicate with any other device directly, without requiring the services of a centralized host. Nodes do not send traffic to a central host for processing and relay services, but instead communicate directly with the destination system, if possible.

When this is not possible—such as when the two hosts are on separate networks—then the sending device has to locate another device to relay the traffic to the destination system on its behalf. Even in this situation the sending device is still self-deterministic, since it chooses which local device it will send the datagrams to for forwarding.

The process of choosing an intermediate forwarding device is called routing. Whenever a device needs to choose a forwarder, it looks at a local list of available networks and forwarders (called the “routing table”), and decides which interface and forwarder is the most appropriate for the specific datagram that needs to be sent.

As was discussed in Section 2.1.2 earlier in this chapter, the routing table on a system can be built using several different tools. To begin with, most systems build a basic routing table that shows the available network interfaces and the networks they are attached to. This information can then be supplemented with manual entries that identify specific forwarders for specific networks and hosts, or a simple “default ...

Get Internet Core Protocols: The Definitive Guide now with the O’Reilly learning platform.

O’Reilly members experience books, live events, courses curated by job role, and more from O’Reilly and nearly 200 top publishers.