Illustrated TCP/IP
by Matthew G. Naugle Wiley Computer Publishing, John Wiley & Sons, Inc. ISBN: 0471196568 Pub Date: 11/01/98 |
Previous | Table of Contents | Next |
RIP Operational Types
|
There are two types of RIP packets that traverse a network (indicated by the command field, shown next): one to request information and the other to give information (a response packet). A response packet is generated for a request packet and is used for periodic RIP updates. Most RIP packets that traverse a local network will be the periodic RIP table updates. Does RIP operate in a workstation? The answer is, yes and no. Some workstations implement RIP. Windows NT and Unix both use RIP; however, most simple workstations such as Windows 95 do not (they have a default gateway). If a workstation does implement RIP, it is usually in what is known as passive mode, which means it can receive and process updates, but cannot respond to RIP requests or broadcast its table.
In passive mode, RIP listens only for RIP updates. (It may build its own tables or it may not. If it does, it will not broadcast these tables.) It will build a table so that it will not have to request information from other routers on the network. Passive end is used for nonrouting network stations. These devices have no reason to broadcast updates, but have every reason to listen for them. Today, most DOS PC computers will use a concept of a default gateway, explained later. Even Windows 95 uses a default gateway if prompted. It can build a routing table, but Windows 95 is not RIP-enabled.
The RIP passive protocol allows the host to maintain a table of the shortest routes to a network and designates which router to send the packets to. This consumes a considerable amount of RAM for both the table and the algorithm. Without it, TCP/IP requires a default gateway entry, which specifies that when a packet is destined for a remote network, the host must submit the packet to a specified gateway for processing, even if this gateway is not the shortest path to that network. Passive implementations add no overhead to the network, for they listen only to routing table updates that are on the network. Without passive RIP, these devices have to maintain their own tables or implement a default route.
Most workstations do not invoke active versions of the RIP protocol They do not build tables and keep track of networks. To communicate with a router, workstations generally use their default gateway parameter. This is for simplicity. Higher-powered workstations, such as Sun SPARC workstations, can build and maintain routing tables. However the early implementations of ICP/IP were not powerful and required a simple method.Remember, RIP packets do not leave their local network. All participants in the RIP protocol (for example, routers) will receive the packet, update their tables if necessary, and then discard the packet. They will compute the reachability of networks based on adding a cost (usually 1) to the just-received tables or count entry, and then broadcast their tables out their ports (usually being mindful of a protocol named split horizon, which is explained a little later).
Previous | Table of Contents | Next |