Distance Vector Routing Protocols Distance Vector Routing

Distance Vector Routing Updates Routing updates occur periodically or when the topology of the network changes. Updates proceed from router to router. Each router sends its entire routing table to each of its adjacent neighbors. Routing tables include information about:  Total path cost as defined by the metrics  Logical address of the first router on the path to each network contained in the table

Distance Vector Update Process

Routing Loops Introduction Routing loops can occur when routing tables are inconsistent due to slow convergence in a changing network.

Converged Network Just before the failure of the /24 network, all routers have consistent knowledge and correct routing tables. The network is said to have converged. For RTC, the preferred path to the /24 network is through RTB, and the distance from RTC to the /24 network is 3.

Converged Network Example RTE#show ip route C /24is directly connected, FastEthernet1/0 RTA#show ip route R /24[120/1] via , FastEthernet0/0 RTB#show ip route R /24[120/2] via , Serial2/0 RTD#show ip route R /24[120/2] via , Serial2/0 RTC#show ip route R /24[120/3] via , Serial2/0

Network Failure After the /24 network fails and an update is scheduled to be sent, RTE sends an update to RTA. RTA stops routing packets to the /24 network, but RTB, RTC, and RTD continue to do so, because they have not yet been informed of the failure. When RTA sends out its normal update, RTB and RTD stop routing to the /24 network. However, RTC has not received an update. For RTC, the /24 network can still be reached through RTB.

Network Failure Example RTE#show ip route No route to /24 RTA#show ip route No route to /24 RTB#show ip route No route to /24 RTD#show ip route No route to /24 RTC#show ip route R /24[120/3] via , Serial2/0

The Routing Loop Now RTC sends a periodic update to RTD, which indicates a path to the /24 network through of RTB. RTD changes its routing table to reflect this incorrect information, and sends the information to RTA. RTA sends the information to RTB and RTE, and the process continues. Any packet destined for the /24 network will now loop from RTC to RTB to RTA to RTD and back again to RTC.

The Routing Loop Example RTE#show ip route C /24is directly connected, FastEthernet1/0 RTA#show ip route R /24[120/5] via , Serial3/0 RTB#show ip route R /24[120/6] via , Serial2/0 RTD#show ip route R /24[120/4] via , Serial3/0 RTC#show ip route R /24[120/3] via , Serial2/0

Looping Forever Packets will loop continuously around the network, despite the /24 network being down. Each time a router handles a packet it increases the packet ’ s hop count value.  Counting to infinity

Combating Routing Loops A number of dynamic routing protocol measures can prevent routing loops from occurring or limit the problems they cause.  Defining a maximum hop count  Split horizon  Route poisoning  Triggered updates  Holddown timers

Defining Infinity Distance-vector protocols define infinity as a specific maximum number of hops.  This number usually refers to a maximum hop count.  15 for RIP, 255 for IGRP The routing protocol permits the routing loop to continue until the hop count reaches ‘ infinity ’. When the maximum hop count is reached, the /24 network is considered unreachable.

Split Horizon Using split horizon, a router will not advertise a route update to the router that originally informed it of the route.

Split Horizon Example (1) RTA, RTB, and RTD exchange updates indicating the network is down.  RTC has not received the message, yet.

Split Horizon Example (2) Now, RTC sends a periodic update to RTD, which indicates a path to the /24 network through of RTB. RTD changes its routing table to reflect this incorrect information.  RTD will not send the updated information to RTA, as RTD originally learned the /24 network was down through RTA.

Split Horizon Results RTE#show ip route No route to /24 RTA#show ip route No route to /24 RTB#show ip route No route to /24 RTD#show ip route R /24[120/4] via , Serial3/0 RTC#show ip route R /24[120/3] via , Serial2/0

Split Horizon Completion Eventually, RTB will send a periodic update to RTC, indicating the /24 network is unavailable. Then RTC will update its routing table and send an update to RTD.  Then, RTD will clear its route, again.

Route Poisoning Poisoned routes are advertised with a distance higher than the maximum hop count.  For RIP, poisoned routes are advertised as being 16 hops away.  Indicates an unreachable network By itself, route poisoning does little.  When used with triggered updates, route poisoning speeds-up convergence time, by immediately notifying other routers when a network is down.

Triggered Updates By default a router running RIP advertises its routing table very 30 seconds.  Every 90 seconds for routers running IGRP Using triggered updates, when a route fails, an update is sent immediately. The router that detects a topology change immediately sends an update message to neighbor routers that, in turn, send triggered updates to their neighbors.

Holddown Timers When a router receives an update indicating a route is down, it marks the route as inaccessible and starts a hold- time timer. If the holddown timer expires, the route is completely removed from the routing table.

Holddown Timer Changes If at any time before the holddown timer expires an update is received from:  The same neighbor indicating that the network is again accessible, the router marks the network as accessible and removes the holddown timer.  A different neighbor with a better metric than originally recorded for the network, the router marks the network as accessible and removes the holddown timer.  A different neighbor with a poorer metric: The update is ignored. This allows more time for the knowledge of a disruptive change (routes frequently flapping from up to down) to propagate through the entire network.

Holddown Timers Illustrated