Routing Concepts and Distance Vector Protocols, Part—II : Dynamic Routing(Source:Network Professional's Advaned Internetworking Guide, John WileyPublications)
Two versions of RIP, version 1 and version 2, work for IPv4 addressing. The primarydifference between the versions is that version 1 is a classful routing protocol and version 2is a classless routing protocol.
Distance Vector Routing in General:
operate only a small amount of information about thenetwork; need basic information about destination networks. Distance vector protocols mustknow
a specific distance and direction (or vector) to each destination network
, hence thename of the protocols.
Distance vector routing protocols keep track of changes to the internetwork by
broadcasting periodic routing updates out all active interfaces.
Each broadcast includesthe complete routing table.
This can work okay, but the amount of CPU process and link bandwidth can be morethan you might want.
The slow convergence of distance vector routing protocols can result in inconsistentrouting tables and routing loops. Now for a
to occur, a sequence of eventswould have to happen in perfect timing. If BrusRtr4 had sent its last update 1 secondbefore network E went down, then BrusRtr4 is not due to send another update for 29more seconds. (RIP’s update timer is 30 seconds.)
RIP uses a very simple metric of hop count--simply the number of routers thatmust be traversed to get to a destination network. For a network that is directly connected,the hop count is 0.
Distance Vector Routing Protocol Looping Avoidance Mechanisms:
It’s caused by gossip(broadcasts) and wrong information being communicated then propagated throughout theinternetwork.
One way of solving this problem is to define a
maximum hop count
. RIP permits a hopcount of up to 15, so any network that requires 16 hops is deemed unreachable.
It is never useful to send routing information back in the direction fromwhich it was learned. In other words, the routing protocol identifies the interface anetwork route was learned on and won’t advertise the route back out that same interface.
Poisoning the route, or sending updates for the route with hop count16, to a downed network keeps other routers from being susceptible to incorrect updates.When a router receives a poisoned route, it sends an update, called a
,back to the notifying router. This breaks split horizon rule.
Holddowns prevent routes from changing too rapidly by allowing time foreither the downed route to come back up or the network to stabilize somewhat beforechanging to the next best route.
Link State Protocols:
Link state routing protocols are classless routing protocols. Again, tobe a classless routing protocol, the subnet mask information is carried with the routingupdate so that all of the neighbouring routers know how large the advertised netwok is.
Except maintaining a routing table with the destination routes in it Link state routingprotocols maintain two additional tables--a
maintained through the use of
exchanged by all routersto determine what other routers are available for exchange routing data. All routers thatcan share routing data are stored in the neighbor table.
is built and maintained through the use of
link state advertisements
(LSA) or link state packets (LSP), depending on the protocol. The table contains a listingfor every destination network for every neighbor that the router can talk to.