moving on to how to prevent these routing loops from happening in the first place.
Well, the first method deployed by this inspector routing protocol is having maximum hop counts
four packets or routing updates.
So, as you saw when that loop was happening in the counting to infinity, problem was happening.
Rip deploys a maximum hop count of 15
so a routing update can go a maximum of 15 router hops.
The 16th router kills the packet, so when the maximum hop count is reached, the packet is killed, so it doesnt loop around the network endlessly.
The 2nd 1 is split arising.
Split horizon basically saves. Let's say this is Router one
router to and again. Network 10
Network 10 is available to Router one through a router to writer, too, is sending updates about Network 10 going out Serial 00 interface
So the rule of sweet horizon basically say's if I send an update about a network outbound on an interface,
I will not accept an inbound a bit about the same network
coming in on that interface.
if Route or two cents an update about Network 10 out of the serial 00 interface router to will not accept an obit about Network 10 coming in on the same interface. And it makes sense the only way network Chen would arrive on
cereals, you know, zero on router to serial 00 interface.
If there was a loop here somewhere downstream on the a bit looped around is now on its way back.
So the ruler, split arising, stays once again. If I send an outbound update out and interface about a certain network, I am not going to accept an update in bound on the same interface about the same network
moving on the rule of route poisoning.
The rule of route poisoning basically
a down route with an infinite metric.
16 is infinite. The maximum hop count for route information protocol or rip, is set to 15.
So if this network for any reason goes down, this link fails router to will immediately send an update to Router one
telling router one. That Network 10 now has a hot count off 16 or an infinite hot cop count, and you can never get to an infinitely far away destination. So Router one won't even bother trying to route to Network 10.
The 4th 1 is Route Invalid Timer.
Let's say there is another router here
connecting router one, which is Router three.
And once again, Network 10 is available to Router one and Router three through router to
let's say this interference, which is cereal 00 again on Rather one has a malfunction
and router One stop sending updates to Router three about network 10.
Well, Router three will wait the length of the route Invalid timer,
which for rip is 182nd.
And then, if it does not hear about
network 10 in 180 seconds, Router three will mark that network as possibly down and stop routing to it.
the route invalid timer is the length of time a router will wait to hear a routing update from an upstream router before declaring that route to be possibly down and stopping any routing towards it.
The next one is hold down, timer.
when the invalid timer expires, the whole down timer is started.
The whole down time of four. Rip is 180 seconds.
So when the invalid timers expires
and the route is considered possibly down the hole down timer starts
a route. When it's put in hold down.
When a router puts air out and hold down, it will not accept any a bit about that route for that length of time with rip being 180 seconds.
well back in the olden days? Router information, router interfaces when they went down and when they came back up, they used to flap.
They used to go down, come up, go down, come up, go down, come up before stabilizing.
So this was a mechanism
for the router not to continually add and delete that route from the routing table as the upstream interface came up and down, came up and went back down. So router three, in that case would say, Okay, hold on.
This network was down for a while. Now it's back up. But I'm not going to accept
any new updates for this network for the length of the whole down timer, because I'm not sure whether you're stable yet,
have you come up and you decide to stay up for a period of 180 seconds. Then I will start accepting updates about this network because I don't want to constantly ad network 10 and then take Network 10 away from my routing table because when I'm doing that, I'm not actually routing
and packets could back up.
So when they're out, invalid timer expires, which is 180 seconds. The whole *** timer is started, which is
now, while the router is in.
for network 10 is in hold down.
Router three will send a poisoned a big upstream back to router one saying, Hey, I haven't heard about Network 10
for 180 seconds. I am putting this route and hold down, and I'm not accepting any updates about Network 10 for a period of 180 seconds and while and hold down, I'm also gonna tell Router one. Hey, get rid of this route from your routing table because I haven't heard about it. 480 seconds and it's probably down.
That process is called poison reverse. So basically, Router three will send
an update about Metro 10 back to Router one with the hop count of 16 which in rip is infinite. So advertising a route back upstream
advertising. A poison route back upstream towards the region of the network
with an infinite metric
or an unreachable metric is called route poisoning.
The last one is route flush timer. This is the time at which the route is completely removed from the I P routing table for rip. That is 120 seconds
now. One last note on rip.
Right now we're doing ripped version one or simply called rip or route Information protocol
ripped. Version one is class full, which means it only understands class A Class B and classy addressing.
Also, that means that sub net mask information is not carried within the routing update.
Okay, you look at the first octet and you know it's class, eh? You look at this first doctor, you know, it's Class B or classy.
Okay. And us human mash a mask off either slash eight, staff 16 or slash 24.
This concludes our route information protocol lecture, and now we'll get to the lab