In order for routers to forward packets to remote networks, they must know how to reach them.  There are two options: static or dynamic routes.

Static Routes

Static routes are manually configured on each router.  They are used for a couple of reasons:

• where there is only a single path to a network (a.k.a. stub network)
• when connecting to an ISP and configuring it as a default (static) route

There are a number of problems with implementing static routes network-wide.  Some include:

• failure to scale well
• does not automatically react/recover to changes in the network
• tedious to configure for large networks (see point 1)

To configure a static route:

R1(conf)# ip route prefix mask address|interface [distance]

The prefix and mask is the destination network and subnet mask.  You can use an address to define the IP address of the next hop towards the destination network or specify a local router interface that the router will use to send traffic out to the destination network.  The optional distance keyword can be used to manually define the administrative distance for the route.

Static Default Routes

One of the most common uses of static routes is for creating a default route.  There are often cases when you want to forward packets that are not defined in a specific route out an interface or towards another router.  A common example is when connecting to an ISP.  If traffic is destined for an address range not defined within your organization (i.e. your coworker’s Facebook updates), then it makes sense to configure a default route towards your ISP or other organization.

To configure a static default route:

R1(conf)# ip route 0.0.0.0 0.0.0.0 address|interface

Floating Static Routes

There are some circumstances when it makes sense to use a static route as a backup to a dynamic routing protocol.  In order for this to work, however, the default administrative distance value on the static route must be raised so it will have a lower priority than the dynamic routing protocol (see administrative distance section below).

Dynamic Routing

Dynamic routing protocols can dynamically respond to changes in the network.  The routing protocol is configured on each router and the routers learn about both each other and remote networks.

Examples of modern dynamic routing protocols include:

• RIP v1,2 (ok, maybe this isn’t very “modern”)
• EIGRP
• IS-IS
• OSPF
• BGP

Distance Vector vs. Link-State

Distance Vector
When routers run a distance vector dynamic routing protocol, they periodically send information about their known routes to their connected neighbors.  This is how the router knows whether changes have been made to the network.  They compare their routing table against the information they receive from their neighbors – if it matches, they’re good.  If not, they update their routing tables to reflect the changes.

RIP is an example of a distance vector routing protocol.

Link State
Link state routing protocols operate differently.  Routers send information about the state of their links to the entire network (or area) that they are a part of.  In this way, each router understands the entire network topology and must run an algorithm every time a network change is announced to recalculate the best routes throughout the network.  This makes link state routing protocols much more processor intensive.

The second major difference in link state routing protocols is that updates are only sent if a change on a router’s link occurs.  This helps keep bandwidth utilization low, unlike distance vector protocols which send out reoccurring updates regardless if a change has occurred.

OSPF and IS-IS are examples of link state routing protocols.

Advanced Distance Vector
Advanced distance vector is the title Cisco gives to EIGRP, which borrows the best attributes of both distance vector and link state designs.  EIGRP does not send periodic route information; instead it sends updates only when changes occur (like link state protocols).  Also, EIGRP forms neighbor relationships with its directly connected peers and only updates them – not the entire network (like distance vector protocols).

Classful Concepts

IP routing protocols are either classful or classless and that determines how they present route information.

Classful
Classful routing protocols (like RIPv1) do not include the subnet mask in routing updates.  When an update is sent, the packet contains only the major network information depending on whether it is a class A,B, or C address.

For example, a route to network 172.16.10.0/24 would be advertised as 172.16.0.0/16 because its classful boundary is a class B address.  Obviously if you have broken your major network boundaries up into smaller subnets that are more granular than the major classful boundaries, this will not work well and that’s the reason almost all modern routing protocols are classless.

Classeless
Classless routing protocols (like RIPv2, EIGRP, OSPF, IS-IS, and BGP) include the subnet mask in routing updates allowing for VLSM support and supernetting.

Administrative Distance

Routers need a way of determining which path to use to a destination network if two or more routing protocols are in use and both advertise a route.  Administrative distance is Cisco’s answer.  Cisco has assigned an administrative distance (AD) to each routing protocol that outlines which protocol a router will prefer.  The AD values can be between 0 and 255 with the lowest values being used for routing.

The table below shows the default AD values

For example, if router R1 receives a route to network 10.10.10.200.0 from both EIGRP and OSPF, the router will compare the administrative distance of the EIGRP-learned route (90), to that of OSPF (110).  The router will then add EIGRP’s route to the routing table because its AD is lower (90 < 110).

Summary