Tip:
Highlight text to annotate it
X
Now you know how a router uses its routing table and how to configure a static route.
Are there any disadvantages to static routes? Let me show you an example:
Here I have many routers and a lot of networks. If I want to configure full reachability between
the routers then I have to configure a LOT of static routes to make this work and you
don’t have any backups. If a link fails you'll need to edit your static route and
send traffic another direction. In this case would be better to use the dynamic routing.
Dynamic routing is where we use a routing
protocol. Routing protocols will send network information to each other and they will keep
it up-to-date. If there are any changes in the network our routers will update each other
to reflect this new information. If we talk about dynamic routing there are
2 different terms that sound similar but are completely different, Routing protocols and
Routed protocols A routing protocol is used between routers
to exchange routing information and build the routing table.
Routed protocols are the protocols that we are routing…for example IPv4 or IPv6.
We have different types of routing protocols; let me show you this picture:
Let's break this one down: An autonomous system is a collection of routers/networks that belong
to a single administrative domain. Which is the network you are responsible for. Your
company network could be an autonomous system. Your internet service provider has its own
network which is another autonomous system.
Within an autonomous system we run a routing protocol and we call these interior gateway
protocols or IGP. OSPF and EIGRP are IGP‟s and we will discuss those 2 routing protocols
in detail in the upcoming videos.
Between the autonomous systems we also run a routing protocol but we call these external
gateway protocols or EGP. So the whole internet is a bunch of autonomous systems connected
to each other and in between we run an EGP. There is only one routing protocol we use
on the internet which is called BGP or Border Gateway Protocol.
Now we will only talk about the interior gateway protocols and take a close look at OSPF and
EIGRP. Let's focus on the interior gateway protocols
and here I have another picture for you: Here we have an autonomous system with a bunch
of routers. On the left side there's a computer in the 192.168.1.0 network and on the right
side a server in the 192.168.2.0 network. Our computer wants to reach the server on
the right side. It's up to our routers to “route” our IP packets toward the server…the
question is; which path are we going to use to get there?
This depends on the routing protocol you are
using; we have 2 different routing protocols, OSPF and EIGRP. Each of them has a different
view and what they think about, is the shortest path to get from the source to destination.
How do they determine the best path? They do so by using something called metrics.
OSPF and EIGRP each use different metrics and we'll discuss them in detail.
Metrics are what routing protocols use to determine the shortest path to the destination.
OSPF uses something called cost as the metric and it will actually look at the bandwidth
of an interface. OSPF will prefer a 100Mbit link over a 10Mbit link.
EIGRP is the Cisco routing protocol and it can use multiple metrics. It will look at
the bandwidth and delay of an interface and if you want, it can look at the load and reliability
of an interface as well.
The question now, Can we use multiple routing protocols at the same time on a router? Yes
we can…perhaps because your company bought another company that has routers running another
routing protocol and you have to merge the two networks.
We do however get some funky situations like
in the following example: This time we are running two routing protocols
on our network, OSPF and EIGRP. Both routing protocols are giving us routing information:
• EIGRP tells us the router should send
IP packets using the path on the top. • OSPF tells us the router should send IP
packets using the path on the bottom.
What routing information are we going to use? Both? Use OSPF or EIGRP?
The answer is that when two routing protocols are giving us information about the same destination
network we have to make a choice…you can't go left and right at the same time. We need
to look at the administrative distance or AD.
Let me show you the administrative distance
list: The lower administrative distance is the better.
As you see a directly connected route has an AD of 0. This makes sense since there's
nothing better than having it directly connected to your router. A static route has a very
low administrative distance of 1 which also makes sense since this is something you configure
manually. Sometimes you use a static route to “overrule” a routing protocol's decisions.
EIGRP has an administrative distance of 90
which makes sense since it's a Cisco routing protocol. And OSPF has 110. In our example
here, we will use the EIGRP since its AD is 90 that is lower than OSPF which is 110.
Make sure you understand the concept of metrics and administrative distance before you continue.
Interior routing protocols are divided in
three different classes:
• Distance Vector : uses simple algorithms that calculate the distance value between
routers based on hop count. And RIP belongs to this class.
• Link-State : calculates the path between routers based on a complex database. OSPF
belongs to this class. • Hybrid: which uses a combination of distance-vector
and link-state methods that tries to incorporate the advantages of both and minimize their
disadvantages. EIGRP belongs to this class. We will look at the differences of each routing
protocol later. For now Just remember, to which class each routing protocol belongs
to. One last thing to tell you about routing protocols:
Routing protocols can be classful or classless: • Classful routing protocols don't send
the subnet mask along with their updates. • Classless routing protocols send the subnet
mask along with their updates.
Let's take a look at our example here: Here we have three routers and a bunch of
networks. Take a close look at the networks that we have here:
172.16.1.0 /24 172.16.2.0 /24
192.168.12.0 /24 192.168.23.0 /24
• 172.16.1.0 and 172.16.2.0 fall within
the class B range. • 192.168.12.0 and 192.168.23.0 fall within
the class C range.
What subnet mask do class B and C have by default?
• Class B is 255.255.0.0
• Class C is 255.255.255.0 A classful routing protocol will not send
the subnet mask along with the routing update so this is what will happen:
Router 1 and router 3 don't send the subnet mask along with the routing update so it will
advertise the classful network which is 172.16.0.0 in this case. So what happens with router
2? It thinks it can reach the 172.16.0.0 network by sending packets either left or right and
if the metric is equal it will try to load-balance. Obviously this is going to cause problems.
Classless routing protocols advertise the
subnet mask along with their updates: So here as you can see router 1 is now advertising
its 172.16.1.0 network with a subnet mask. Router 3 is advertising its 172.16.2.0 network
with a subnet mask as well.