The current version of OSPF for IPv4 is OSPFv2 (RFC 1247, RFC 2328).
The current version of OSPF for IPv6 is OSPFv3 (RFC 2740).

OSPF has a default administrative distance of 110.
Routing table source code “O”.
Enabled with “router ospf [process-id]” – the process ID does not have to match other OSPF routers.
The “network” command has slightly different syntax. “network [network-address] [wildcard-mask] area [area-id]

The wildcard mask is the inverse of the subnet mask. The area-id should be set to 0.

OSPF does not use a transport layer protocol – packets are sent directly over IP. By default, hello packets are sent every ten seconds, or thirty seconds on NBMA segments.

The Dead interval is the period of time an OSPF router will wait before terminating adjacency with a neighbour. The Dead interval is four times the Hello interval, by default. (40 / 120 seconds).

“show ip ospf neighbors” can be used to verify OSPF adjacencies.

The OSPF router ID is used to uniquely identify each router in the OSPF routing domain. Cisco routers derive the router ID based on three criteria and with the following precedence:

1. Use the IP address configured with the OSPF router-id command.
2. If the router-id is not configured, the router chooses highest IP address of any of its loopback interfaces.
3. If no loopback interfaces are configured, the router chooses highest active IP address of any of its physical interfaces.

RFC 2328 does not specify which values should be used to determine the cost. Cisco IOS uses the cumulative bandwidths of the outgoing interfaces from the router to the destination network as the cost value.

Multiaccess networks can create two challenges for OSPF regarding the flooding of LSAs, including the creation of multiple adjacencies – one adjacency for every pair of routers, and extensive flooding of LSAs (Link-State Advertisements). OSPF elects a DR (Designated Router) to act as collection and distribution point for LSAs sent and received in the multiaccess network. A BDR (Backup Designated Router) is elected to take over the role of the DR should the DR fail. All other routers are known as DROthers. All routers send their LSAs to the DR, which then floods the LSA to all other routers in the multiaccess network.

The router with the highest router ID is the DR, and the router with the second highest router ID is the BDR. This can be superseded by the ip ospf priority command on that interface. By default, the ip ospf priority is “1″ on all multiaccess interfaces. If a router is configured with a new priority value, the router with the highest priority value is the DR, and next-highest the BDR. A priority value of “0″ means the router is ineligible to become the DR or BDR.

A default route is propagated in OSPF similar to that of RIP. The OSPF router mode command, default-information originate is used to propagate a static default route.

The show ip protocols command is used to verify important OSPF configuration information, including the OSPF process ID, the router ID and the networks the router is advertising.

Built around Edsger Dijkstra’s shortest path first (SPF) algorithm. There are two link-state routing protocols for IP: OSPF (Open Shortest Path First) and IS-IS (Intermediate-System-to-Intermediate-System).

The link-state process can be summarised as follows:

1. Each router learns about its own directly connected networks.
2. Each router is responsible for saying hello to its neighbours on directly connected networks.
3. Each router builds a Link-State Packet (LSP) containing the state of each directly connected link.
4. Each router floods the LSP to all neighbours, who then store all LSPs received in a database.
5. Each router uses the database to construct a complete map of the topology and computes the best path to each destination network.

A link is an interface on the router. A link-state is the information about that interface including its IP address and subnet mask, the type of network, the cost associated with the link, and any neighbour routers on that link.
Each router determines its own link-states and floods the information to all other routers in the area. As a result, each router builds a link-state database (LSDB) containing the link-state information from all other routers. Each router will have identical LSDBs. Using the information in the LSDB, each router will run the SPF algorithm. The SPF algorithm will create an SPF tree, with the router at the root of the tree. As each link is connected to other links, the SPF tree is created. Once the SPF tree is completed, the router can determine on its own the best path to each network in the tree. This best path information is then stored in the router’s routing table.

Link-state routing protocols build a local topology map of the network that allows each router to determine the best path to a given network. A new LSP is sent only when there is a change in the topology. When a link is added, removed or modified, the router will flood the new LSP to all other routers. When a router receives the new LSP, it will update is LSDB, rerun the SPF algorithm, create a new SPF tree, and update its routing table.

Link-state routing protocols tend to have a faster convergence time than distance vector routing protocols. A notable exception is EIGRP. However, link-state routing protocols do require more memory and processing requirements. This is usually not an issue with today’s newer routers.

There are level 1 and level 2 routes. A level 1 route can be either an ultimate route or a parent route.

Level 1 Ultimate Route:
A route with a subnet mask equal to, or less than the default classful mask of the network; and either a next hop address or an exit interface. For example, a route learned through RIP with the network address of 192.168.1.0 and a /24 network mask is a level 1 ultimate route. These are displayed in the routing table as a single route entry, such as:

R 192.168.1.0/24 [120/1] via 172.16.2.2, 00:00:25, Serial0/0/0

Level 1 Parent Route:
Automatically created when a subnet route is added to the routing table. The subnet route is known as a level 2 child route. The parent route is aheader for level 2 child routes. An example of a parent with a level 2 child route is below:

172.16.0.0/24 is subnetted, 1 subnets
R 172.16.1.0 [120/1] via 172.16.2.1, 00:00:07, Serial0/0/0

The subnet mask of the child routes are displayed in the parent route unless VLSM is used. With VLSM, the parent route displays the default classful mask and the subnet mask is included with the individual VLSM route entries.

Routing Table Lookup Process:
When a packet is received by the router, it looks for te longest match with one of the routes in the routing table. The longest match is the route with the largest number of left-most bits that match between the destination IP address of the packet and the network address of the route in the routing table. The subnet mask associated with the network address in the routing table defines the minimum number of bits that must match for that route to be a match.

Before examining level 2 child routes (subnets) for a match there must first be a match with the level 1 parent route. The classful mask of the parent determines how many bits must match the parent route. If there is a match with the parent route then the child routes will be searched for a match.

When there is a match with the parent route but none of the child routes, then, if the router is using a classful routing behaviour, no other routes will be searched – the packet will be discarded. Classful routing behaviour can be implemented using the “no ip classless” command.

Beginning with IOS 11.3, classless routing behaviour became the default (it was previously classful by default). If there is a match with a parent route but none of the child routes, the routing table process will continue to search other routes in the routing table including a default route, should one exist. Classless routing behaviour is implemented using the “ip classless” command.

Routes to networks get added to the routing table from various sources including directly connected networks, static routes, classful routing protocols, and classless routing protocols. The lookup process, classful, or classless routing behaviour, is indepentent of the source of the route. A routing table may have routes learned from a classful routing protocol such as RIPv1, but uses classless routing behaviour (“no ip classless”), for the lookup process.

RIPv1 can not support discontiguous networks because it automatically summarises at major network boundaries. A router that receives routing updates from multiple routers advertising the same classful summary route cannot determine which subnets belong to which summary route. This inability leads to unexpected results including misrouted packets.

The default version if RIP is version 1. Entering the command “version 2″ while configuring RIP will cause it to switch to version 2.

Similar to RIPv1, RIPv2 automatically summarises at major network boundaries. However, with RIPv2 automatic summarisation can be disabled with the “no auto-summary” command. Automatic summarisation must be disabled to support discontiguous networks.

The command “debug ip rip” will show RIP updates – including the subnet mask.

The “show ip protocols” command will display that RIPv2 is being sent and received, and also whether or not automatic summarisation is in effect.

Address classes are no longer needed with CIDR – network addresses are determined by subnet mask rather than the value of the first octet. This means that the number of hosts available on a network can be varied to more specifically suit the number of hosts needed.

CIDR allows for a group of major network addresses to be summarised to a single network address with a mask less than the default classful mask. This is called supernetting.

CIDR uses Variable Length Subnet Masks (VLSM) to allocate IP addresses to subnets according to need rather than by class. VLSM allows for subnets to be further divided or subnetted into even smaller subnets. VLSM is essentially subnetting a subnet.

Because of the changes CIDR brings, a classless routing protocol is required (that is, one that includes the subnet mask with the routing update).

Determining the summary route:
1. List the networks in binary format.
2. Count the number of left-most matching bits.
3. Copy the matching bits then add zero bits to the rest of the address.

This can then be used by both static routes and classless routing protocols as the summary route. Classful routing protocols can only summarise routes to the default classful mask.

RIP – Routing Information Protocol (RFC 1058)

Characteristics:

• Hop count is used as the metric for path selection.
• RIP cannot supply a route to a network with a hop count greater than 15.
• Routing updates are, by default, broadcast every 30 seconds.

IGRP – Interior Gateway Routing Protocol

Characteristics:

• The metric for path selection is based on a combination of bandwidth, delay, load, and reliability.
• The default period for routing updates is 90 seconds.
• IGRP is now obsolete, having been replaced by it’s descendant, EIGRP.
• Cisco proprietary.

EIGRP – Enhanced Interior Gateway Routing Protocol

Characteristics:

• Possibility of unequal cost load balancing.
• Uses Diffusing Update Algorithm (DUAL) to calculate the shortest path.
• There are no periodic updates. Routing updates are sent only when there is a change in the topology.

Distance Vector

Means that routes are advertised as vectors of distance and direction. The meaning of distance can change based on the metric used. A router using a distance vector routing protocol knows only the direction in which packets should be forwarded, and how far it is to the destination network.

Common Characteristics

Periodic Updates are sent at regular intervals, whether or not the topology has changed, for both RIP (30 seconds) and IGRP (90 seconds).

Neighbours are routers that share a link and are configured with the same routing protocol. The router is only aware of the addresses of its own interfaces and any neighbour’s addresses. There is no broader knowledge of network topology.

Broadcast Updates are sent to 255.255.255.255. Neighbouring routers will process updates. Other devices process the update up to Layer 3 but then discard it. Some distance vector routing protocols use multicast instead of broadcast.

Entire Routing Table Updates are sent periodically to all neighbours. Information that is relevant is found by processing the entire update and then discarding unneeded information. EIGRP does not send such updates.