Dynamic Routing

we remember on the router there can be static routing that is configured it's a common misconception that boy we would never do static routing why would we do static routing when we can have the routers do dynamic routing well keep in mind if you have what is called a stub network let's say we have router one here and this router connects into an area of the network and behind this router are just a couple of segments that have no other way to get to the rest of the world except through our one we call this a stub network and with this stub network we can go ahead and configure a simple default static route to get everyone to the rest of the world so yeah dynamic routing is wonderful but we are still going to have plenty of environments in which we have static route and used dynamic routing of course we allow the routers to exchange information in order to fully populate those routing tables that we utilize dynamic routing protocols are indeed going to be categorized into two categories and one of those categories is called interior gateway protocols and these are protocols like EW grp and OSPF that run inside our organization we take our organization and we may indeed take prefixes and advertise them out to the Internet and this is done with what is called exterior gateway routing protocols and there's only one of those in use today and it's called BGP so there are several options for interior gateway protocols organization and then there's only one option for exterior gateway protocols and that of course is the border gateway protocol so what are outers tasked with well they're tasked with discovering remote networks maintaining a routing table of that information choosing what is the best path to get to a destination when there are multiple paths and if one of the paths are both of the paths go down what path do we take discovering a new path to the destination now earlier I pointed out to you that we have our interior gateway protocols and we have our exterior gateway protocols and what really determines if we are going to use an e GP or an AI GP is whether or not we're in the same autonomous system yeah the term autonomous system is used to describe a particular network and notice there are autonomous system IDs which will identify organizations for BGP to be able to move traffic between them you know it's interesting so these like EW GRP OSPF rip these are our protocols of our company right so their protocols of our company BGP is the protocol of ISPs and the internet now when you say that you say all right well if ISPs are the only ones that are going to run BGP why would I even care if I don't work for an Internet service provider well keep in mind huge companies like Cisco they are going to go ahead and run BGP because they want to get through the ISPs the information to their biggest customers in the most efficient manner possible now we don't focus on Beach in this class because this is not a class targeted at you know service providers in supporting really big information environments what CCNA focus is on is interior gateway protocols that we are going to run inside of our organization now the first generation technology that came along the first generation technology that came along was distance vector routing protocols in fact it was rip version one that first came along and rip version one it was true distance vector in that it just looked at how many hops away a particular destination was and it looked at the direction also known as vector you can think of airplane here with the famous scene what's our vector victor and then to further complicate things one of the pilots was named Roger so they would say Roger Roger and then I think there was Clarence Clarence well anyways you get the idea vector is Direction so rip said look we're just going to measure things by how many router hops away and in what direction you are direction would be dictated by the particular interface out which you reach that particular destination distance vector routing protocols were improved a bit when rip version to hit so rip was re-engineered into a version two it was still distance vector but as you might guess rip version 2 did feature some improvements over it version 1 then link state routing protocols were invented this is really our second generation technology and OSPF and is is were in the second generation of improved operation then Cisco came along and said we're gonna take the best features of distance-vector and we're going to marry them with the best features of link-state and we're gonna create an advanced distance vector routing protocol called AIG RP enhanced interior gateway routing protocol do you know that as of this month of this year February 2013 cisco has opened AIG RP up to the public that's right it used to be a proprietary Cisco routing protocol and for purposes of your exam I want you to remember that it is Cisco proprietary but as of this month that is no longer true cisco has opened up ew GRP for the entire world to utilize and even modify on their particular own equipment now one of the important distinctions between all these interior gateway routing protocols is how will they figure out what is the best path now we term this the metric yeah one routing protocol will use a specific metric or measurement when it is trying to figure out the best path compared to another routing protocol for example rip will utilize a metric of a hop count yep how many hops away is a particular destination OSPF will use a metric of cost and cost is based on bandwidth so OSPF would look at this and say well I'm gonna choose this path because T 1 links are superior to 256 K links and then there's AI GRP which can use all four of these if you want but by two we'll use bandwidth and delay values as its metric to determine what is the best path rip starts to look a little scary to us doesn't it because if it uses simple hop count for a metric it would consider this the way to go and this is so much more pathetic of a path so yeah rip really does look scary to us in its simplistic operation now what happens if your routing protocol if your router excuse me is running multiple interior gateway protocols what if it's running rip and it's running EW GRP Wow and you know I've seen this slide a million times and I just noticed that there's a typo there that's amazing so that's e aí GRP of course well if your router let's say router a is running both of them and both of them say hey I know how to get to location X who is router a going to listen to what routing protocol should it listen to should it listen listen to a I GRP or should it listen to rip as far as how to get to destination X well the router uses a value called administrative distance in order to solve this question the lower the administrative distance of the routing protocol the more believable it is yeah great job so it's like a golf score so e IG RP is believed or listened to by router a and the rip information is not used because of the worst administrative distance score or a worser or lesser administrative distance score so pretty interesting now do you think Cisco based these scores on sophistication yes indeed that's exactly what they did they said look II I gr P is just better in fact I hate to pick on rip but it's better in every way then rip so sure enough ew gr p gets a lower administrative distance score and it is the routing protocol that is believed when it comes to getting to a particular destination out there on the network now distance-vector routing protocols we said they were the first generation technology as evidenced by rip version one and we said that this first generation technology was very simplistic it would have a vector Hey in which direction are we going to go and it would have a distance a hop count that it would measure something else the distance vector routing protocols do as we can clearly see in this awesome graphic is that they periodically take their routing tables and just send them back and forth to each other now this looks pretty bad right this looks pretty darn bad and think about this fact this happens regardless of the stability of the network oh my goodness so your network is completely stable and your devices are going to be periodically for no reason whatsoever sending their routing tables back and forth like this a lot of people say that rip is an extremely chatty protocol yeah I'll say rip is an extremely chatty protocol all right it is sending its entire routing table to all of its neighbors periodically even if there is no change whatsoever in the network infrastructure so the routing tables have the destination prefixes and they have a simple hop count for the metric in the case of rip let's analyze this 10 dot 4 dot 0 dot 0 is over here and from router a's perspective that is choo hops away that's the distance and what's the vector victor well that is the serial 0 interface so here's the one hop here's the 2 hops away and go the serial 0 interface direction now you see router a assemble that information by listening to what router be told it distance vector routing protocols are what we call routing by rumor yeah this device thinks it knows how to get to a particular destination based on the information that router B gave it this is indeed the concept of routing by rumor folks and you know what it reminds me of it reminds me of a game I played in elementary school the teacher gathered us all up and got us in a big circle and then that teacher had one student whisper a phrase to the other student in the circle so the teacher said the phrase is next Tuesday the starry sky will be beautiful that student whispered it to the next student that student whispered it to the next student it was amazing by the time it went all 20 by the time it was whispered to all 20 of us and the statement got back around to the teacher it was completely different from the original statement I mean completely it was like on January 1st it'll be sunny and partly cloudy when the original statement was next Tuesday night it will be a started beautiful night or whatever right I mean it was completely different this distortion of the truth reminds me of exactly what can happen in the distance vector routing environment bad information can make its way in here is past to here and believed is past to here and believed and can corrupt the environment this is the exact opposite of what link state routing protocols do like OSPF OSPF is going to not rely on the information given to it by other neighboring devices it's gonna build its own picture of what the network is based on a database worth of information that is reliable and of course we'll talk about that more later on so updates come in the router processes these updates updates its routing table and sends this information to the next device which updates its information and of course sends it on now we're going to come to what is probably the funnest depiction the most remarkable section of the course at least for today and that is I want to show you the nightmare that can happen with a distance vector protocol I want to show you that corruption of information just how bad it can be and then we'll show you fixes that were engineered for this problem so let's look at how bad things can get in a distance-vector environment you're gonna love this and I really want to send a big THANK YOU out to the graphics department here at Stormwind comm as you've noticed as we've moved throughout this course the graphics department is unmatched compared to any other training organization that I've ever seen so big thank you to them all right watch this so here we are the happy little Network we will not be happy everybody's fine everybody can reach everybody all of the information in the routing tables is valid you could go in and you could do it line by line and check it just trust me everything you see here is perfectly valid well watch what happens switch router C loses its link for the 10-4 Network see that look at the cable it's frayed it's broken there's a broken cable right there so what it does is in its own routing table it says okay I can't get to 10.4 now the first issue with these first-generation distance-vector routing protocols is that due to a slow convergence router a thinks it can still get there and router B thinks it can still get there wherever the entry is there it is so these devices are still foolishly sending traffic through the environment so first problem slow convergence yuck now watch this based on our distance vector routing protocols worked this device is going to say its routing table right over to see and notice it has an entry for 10.4 that says hey it's out my cereal one interface and I'm one hop away so router C says oh great I now know how to get to ten for its cereal zero is the direction and it's two hops away uh-oh watch what happens this device sends an update to router B router B goes oh my gosh it turns out that I should go out cereal one towards ten for but it's three opps away because router C just told me it's to ops away I'm gonna send this to a a says whoa ten four is available and it's four hops away oh boy you see where this is going a updates BB update C and they're all thinking the ten-four is available and they are all incrementing the hop count to get to that particular Network destination the engineers of rip saw this and said we're geniuses we've got a fix and you're not gonna believe the fix they said we'll just to find a maximum hop count that'll take care of it so they said look if this starts happening and we get up to sixteen hops consider the route dead consider it unreachable so the first safety mechanism if you will that they put in place with rip was a maximum hop count how'd they come up with 16 well I imagine some smart engineer types got into a room and said look ripp is for small networks only so it's really inconceivable to us that we would ever have an environment where there'd be more than 16 router hops wide I mean after all this is designed for small networks so they picked that number by the way did you notice something else that could happen in this environment at one point router B thought the direction was to C and router C thought the direction was to be uh-oh this creates a routing loop router B sends the information and data dasi C turns around and sends it to B and this continues over and over and over again be sentenced to C C sends to B and it just continues endlessly so another security mechanism they put in place that made such common sense was called got a really cool name split horizon split horizon says look if I generate an update if I send out an update about a network that I possess I am NOT going to listen to that information coming in from that interface oh this makes such perfect sense this is how we humans work in communication right when I walk up to you and I say Dan you are the smartest person in this class do we have a dad in class hey that worked out great we have to Dan's in class today dan you're the smartest person in this class dan doesn't look at me and go Anthony I'm the smartest person in this class although it might be good for Dan to do that for his confidence that's not how we communicate so split horizon says look if I send an update out an interface I'm not gonna listen to that same update coming in on that interface now there is one exception to split horizon and that's called poison reverse what is poison reverse all about well it says we will accept an update in on an interface if it is indicating that the network is down yeah see when this network fails we poison the metric 216 and say it's not reachable we will listen to that coming in the same interface we sent it so poison prefixes can go and be an exception to split horizon as we can see here poison reverse overrides split horizon we will listen to poisoned prefix information you know what else they did that was brilliant they said look when something fails in the infrastructure instead of letting our network become unstable especially in the case of like an interface going down and up and down and up we call this a flapping interface rip says you know what we're not gonna listen to anything about that prefix for a certain period of time oh what a nice feature it's called the hold down timer and rip just stubbornly says look there may be a problem with this prefix and as opposed to becoming very unstable we're just not gonna listen to this particular information for this prefix for some period pretty cool and this really does help with a situation like you see on the screen where we have a network going down and up down and up down and up and by the way this is what interfaces do this is what interfaces do when they fail interfaces just don't fail they tend to fade away and that means they they go up they go down they go up they go down and that's obviously a bad thing by the way when the network first goes down we want to make sure everybody knows that right away so they invented the triggered update for rip yeah to improve rip they said look if something does happen don't wait to tell everybody tell them right away so folks all of these mechanisms poison reverse hold down timers split arise in a maximum metric all of these forces work together in a distance vector environment like rip to make sure your routing is stable and healthy and happy

Channel: StormWind Studios

Views: 90,103

Rating: 4.888412 out of 5

Keywords: cisco, systems, routing, icnd1, ccent

Id: DT-pmXMCOS8

Length: 27min 59sec (1679 seconds)

Published: Mon Jul 01 2013