Need help with convergence time!!!!(OSPF, RIP)

[Jsixx]

Ok so!

i am testing convergence times between routing protocols at the minute and some things aren’t adding up. 

The network i am using to test these times is a fully mesh network consisting of 4 routers , 2 Computers and no switches so all routers in the network are directly connected to each other. One computer is connected to R1 and one is connected to R3. The direct link between R1 and R3 is a fast Ethernet cable and is always the best route to take. To get the convergence time I stream a contract stream of traffic from computer connected to R1 to the computer connected to R3. I then unplug the link connecting these two routers and then a new route will be worked out. 

The problem is that a RIP router should have to wait 180 seconds to notice that the route was unreachable and then should have to wait until 240 seconds to remove the route and work out a new route. So why is it that my RIP network is converging in under 30 seconds when really it should take at least 240 seconds ???

thank you in advance. All help would be greatly appreciated.

[paul driver]

Hello

I think what you are seeing is because you are tearing down the interface (shutdown)  RIP is immediately sending/receiving an update marking the route inaccessible -16 hop count, Because of this the route will be removed from the rib straight away and wont go through is holddown/invalid/flush times

However if you create an acl and deny that route once it has been learned without shutting down any interface you will see that route do what you expect it to do via the routing protocol timers

[Jsixx]

Oh ok that’s great thank you ! However I didn’t this with Eigrp, OSPF and RIP 

I did think that RIP would be the slowest, OSPF would be the second slowest and that EIGRP would be the fastest and this was the case but why would that have been ? Because I did the same test on all three protocols ???

[Jsixx]

Oh ok that’s great thank you ! However I didn’t this with Eigrp, OSPF and RIP



I did think that RIP would be the slowest, OSPF would be the second slowest and that EIGRP would be the fastest and this was the case but why would that have been ? Because I did the same test on all three protocols ???

[paul driver]

Hello
Again I think what you are seeing is the way the igps respond to a topology change -
OSPF is different to RIP and EIGRP, it updates its Link State Database via exchanging link state updates /requests/acknowledgments regarding change to a route, As such when both routers have completed this update exchange the routers will then calculate their own best path for the updated route.

Eigrp is also different as to the way RIP and OSPF calculate a route change.
A topology change in EIGRP  can be advertised via various eigrp packets towards a routers eigrp peers and as such when this change occurs the receiving routers are able in most cases to calculate the next best shortest destination path for that route straight way due to its own previously calculated feasible successor for that route, However if the router cannot do this then this is when eigrp will try to find a best destination path for this changed route via querying its own neighbors for a best path.

Hello/Dead timers in ospf relates to its ospf adjacent or neighbors peerings as such they don’t carry any routing information they are used to monitor the availability of these peers so if hello isnt received/acknowledged in a timely manner then these peerings will eventually fail via these timers, This can be said also for the Hello/Holdtimes in eigrp for neighbor adjacency's.

[Jsixx]

Also, how does the protocol know that the interface is down without counting down??