ADTRAN Stub Routing Spezifikationen PDF herunterladen (Seite 229)

Command Reference Guide Global Configuration Mode Command Set

Technology Review

The RIP protocol is based on the Bellham-Ford (distance-vector) algorithm. This algorithm provides that a

network will converge to the correct set of shortest routes in a finite amount of time, provided that:

• Gateways continuously update their estimates of routes.

• Updates are not overly delayed and are made on a regular basis.

• The radius of the network is not excessive.

• No further topology changes take place.

RIP is described in RFC 1058 (Version 1) and updated in RFCs 1721, 1722, and 1723 for Version 2. Version 2

includes components that ease compatibility in networks operating with RIP V1.

All advertisements occur on regular intervals (every 30 seconds). Normally, a route that is not updated for 180

seconds is considered dead. If no other update occurs in the next 60 seconds for a new and better route, the

route is flushed after 240 seconds. Consider a connected route (one on a local interface). If the interface fails,

an update is immediately triggered for that route only (advertised with a metric of 16).

Now consider a route that was learned and does not receive an update for 180 seconds. The route is marked

for deletion, and even if it was learned on an interface, a poisoned (metric =16) route should be sent by itself

immediately and during the next two update cycles with the remaining normal split horizon update routes.

Following actual deletion, the poison reverse update ceases. If an update for a learned route is not received for

180 seconds, the route is marked for deletion. At that point, a 120-second garbage collection (GC) timer is

started. During the GC timer, expiration updates are sent with the metric for the timed out route set to 16.

If an attached interface goes down, the associated route is immediately (within the same random five-second

interval) triggered. The next regular update excludes the failed interface. This is the so-called first hand

knowledge rule. If a gateway has first hand knowledge of a route failure (connected interfaces) or

reestablishment, the same action is taken. A triggered update occurs, advertising the route as failed (metric =

16) or up (normal metric) followed by the normal scheduled update.

The assumption here is that if a gateway missed the triggered update, it will eventually learn from another

gateway in the standard convergence process. This conserves bandwidth.

RIP-Related Definitions:

• Route - A description of the path and its cost to a network.

• Gateway - A device that implements all or part of RIP - a router.

• Hop - Metric that provides the integer distance (number of intervening gateways) to a destination

network gateway.

• Advertisement - A broadcast or multicast packet to port 520 that indicates the route for a given

destination network.

• Update - An advertisement sent on a regular 30-second interval including all routes exclusive of those

learned on an interface.

1 2 ... 224 225 226 227 228 229 230 231 232 233 234 ... 567 568

ADTRAN Stub Routing Spezifikationen Seite 229