Simple routing protocols work well in simple networks, but as networks grow and become more complex, engineers may have to change routing protocols. Typically, engineers switch routing protocols gradually, so multiple routing protocols run in the network for varying lengths of time.
Routers can connect networks (called routing domains or autonomous systems chinesedumps that use different routing protocols. For example, router R1 in Figure 4-1 interconnects the EIGRP (Enhanced Interior Gateway Routing Protocol) and OSPF (Open Shortest Path First) protocols in AS1. R1 is also connected to the ISP (Internet Service Provider) using BGP (Border Gateway Protocol). R1 is called a border router (also called an edge router) because it interconnects different autonomous systems.
The problem in this environment is that each routing protocol collects different types of information and responds to topology changes in its own way. For example, OSPF’s metrics are based on link cost, while EIGRP’s metrics are based on combined metrics.
Another problem is that running multiple routing protocols increases the load on the router’s CPU and memory. For example, R1 in Figure 4-1 needs to maintain separate routing, topology, and database tables, and exchange and process routing information at various intervals.
Finally, routing protocols are not designed to interoperate with other protocols. For example, OSPF and EIGRP metrics are incompatible, and exchanging routing information between these two metrics places additional CPU and memory load on the router.
Reasons to run multiple routing protocols
While we would like to be able to run one routing protocol across the entire IP interconnect network, engineers may need to use multiple protocols for routing for a variety of reasons.
When migrating from an old IGP (Interior Gateway Protocol) to a new IGP. Until the new protocol completely replaces the old one, there may be multiple redistribution boundaries in the network. The same thing happens when companies using different routing protocols merge with each other. In environments where multi-vendor routers are deployed. In these environments, engineers can use Cisco-specific routing protocols, such as EIGRP, on the Cisco portion of the network, and common standards-based routing protocols, such as OSPF, to communicate with devices from other vendors. When an engineer wants to use a new protocol, but the host system still needs support for the old routing protocol (for example, a UNIX host-based router running RIP). Some departments do not want to upgrade their routers to support new routing protocols.
Running multiple routing protocols
When running multiple routing protocols, a router can learn routes from different routing sources. If a router learns a specific destination from two different routing domains, the route with the lowest administrative distance (AD) value will be placed in the routing table. Administrative distance is a parameter used by routers to score the trustworthiness of a routing protocol. Each routing protocol is assigned a value called the Administrative Distance, which is prioritized from most trusted to least trusted. This criterion is the first criterion to be considered when routers decide which routing protocol to choose when multiple protocols provide routing information for the same purpose. Compared with other routes in the table, the path with the lowest administrative distance to the destination will be placed in the routing table. Routes with higher administrative distances are rejected.
When supporting complex multiprotocol networks, engineers should deploy routing protocol designs and traffic optimization schemes carefully.
