Interior gateway protocols can be split into two main types:

Important

The difference between Distance Vector and Link State protocols is that in this DV, each of the neighbours tells the router what its distance is to specific networks, and what its distance is to those networks (but not about the existence of other routers)

In LS protocols, each router will pass on all the information they have about the network. This means that each router can in theory make better pathing decisions, but does increase the load on each individual router.

All the IGPs are designed to do the same job (to advertise routes within an organization and determine best paths). In general, an organization will pick one of the IGPs; in general, we should avoid mixing IGPs.

Router commands

R1# show ip protocols
// Displays routing protocol, informatio sources, etc.

We can see the protocol by looking at the running configuration. To filter it out, use the pipe command to define where to start returning the information:

R1# show run | begin rip // to see everything starting with rip section
R1# show run | section rip // to see only the rip section

To see information recieved via the protocol:

R1# show ip rip database

Note

We can configure other protocols, like OSPF, etc. Simply replace “rip” in the above commands with the appropriate protocol.

This will show all the routes that were learned from neighbours. This shows all routes learned; to see the best routes check the routing table:

R1# sh ip route

Distance Vector routing protocols

In this type of protocol, each router will send its directly Connected neighbours a list of all of its known networks, along with its own distance to each of those networks.

Note

Distance vector routing protocols do not advertise the entire network topology.

Each router only knows its own directly connected neighbours, and the lists of networks those neighbours have advertised. In other words, it does not have detailed topology information beyond its directly connected neighbours.

Note

Distance Vector protocols are often called “Routing by rumour”

Types of distance vector protocols

In this type of protocol, each router describes itself and its interfaces to its directly connected neighbours. The information is then passed unchanged from one router to another; this means that every router ends up learning the full network topology, including every single router and what interfaces they connect to.

Choosing a protocol

Choosing a routing protocol

It is best practice to choose a specific Interior gateway protocols (IGPs), rather than run several within the same network. This means we have to decide which protocol to use. The most common ones are Enhanced Interior Gateway Routing Protocol (EIGRP) and Open Shortest Path First (OSPF).

OSPF vs EIGRP vs RIP

RIP has scalability limitations so it is not typically used in production. As a result, IGP choice for most companies is either EIGRP or OSPF; from those two, OSPF is the more commonly used, since it supports large networks and has always been an open standard (unlike EIGRP which is recently open, but has limited support on non-Cisco equipment).

Although EIGRP can be simpler to implement and maintain, the wide compatibility of OSPF makes it a much more common choice.

Why RIP? Why EIGRP? Why OSPF?Why IS-IS?

Link to original

Equal Cost Multi Path (ECMP)

ECMP allows routers to place multiple routes in the Routing Table if they have an equal Router Metrics cost. By using ECMP, routers will Load Balancing outbound traffic between the different paths.

Note that all Interior gateway protocols (IGPs) will perform ECMP by default, but Enhanced Interior Gateway Routing Protocol (EIGRP) is the only protocol of UnEqual Cost Multi Path.

Note

Load balancing can also be achieved with static routes: simply input both routes in the table.

Link to original

IGP adjacencies

Interior gateway protocols (IGPs) are configured under global configuration and then enabled on a specific interface. When this happens, the router will search for other routers on that specific interface that are also running the same protocol - this is accomplished by sending out and listening for “hello packets”. Whenever there is a match, they will form an adjacency with each other and exchange routing information.

Modern routing protocols use Multicast for the hello packets, which is more efficient than broadcasts used in older protocols.

Note

If we don’t include an interface as part of the routing protocol the router will not attempt to form adjacency on that interface. Additionally, it will not advertise the routes on the excluded interface to other routers on enabled interfaces. To allow an IP subnet to be included, we must use Passive Interfaces.

Link to original