Section 23.13: Routing (Frame 6)                     [prev][home][next]

There are two basic approaches to constructing these routing tables. The static method is to write them all once and for all, based on careful consideration of how far away the IMPs are and the speed of the lines between them. However, traffic patterns may vary and certain lines may become clogged due to overuse. In this case, dynamic methods of maintaining and changing the routing tables are better, which is the approach that ARPANET and now the Internet use.

Every so often, a router (they are no longer called IMPs in the modern Internet) recalculates its routing table by averaging how many packets have gone through a particular line to an attached router. If a certain line is being used too much, then some traffic that was formerly directed through it is rerouted to another. Also, if a line goes down due to some technical problem then the routing tables have to be rewritten. In order to determine where those downed lines are, routers typically send a steady stream of AYA packets: Are You Alive? The router at the other end of the line, upon receiving this, will send back IAA, "I am Alive!" Sometimes it compares timestamps to measure how far out of synch clocks are and how long it took for the packet to arrive. This kind of polling by packets is called pinging because it resembles sonar probing of the ocean floor.