Example of Distance Vector Routing 1 - Georgia Tech - Network Implementation
Udacity
https://www.youtube.com/watch?v=00AAnwgl2DI

Transkript (automatisch erstellt)

0:00 Let's suppose that we have a three node network with the costs on the edges as shown. Initially, each node has a single distance
0:08 vector representing the shortest path cost to each other incident node in the graph. For example, the
0:16 distance between x and x is obviously zero. And the shortest known distance between x and
0:21 y, from x's perspective is one, the direct path. Similarly, the shortest known distance between x and z
0:28 to x at the outset is five because all it knows is the direct path. Note that a
0:33 shorter path between x and z exists via y, but x simply doesn't know about it yet. Now
0:38 in distance vector routing, every node send its vectors to every other adjacent node. And each node then
0:45 updates its routing table according to the Bellman-Ford equation. Let's look at what happens when node x learns of y's distance vectors. Well in
0:55 this case, the distance from x to z will be computed as the minimum of the sums of all distances to z through any
1:04 intermediate node. So the cost between x and y is one, and the distance between y and z as discovered by y's distance vector
1:15 is two. Therefore, x can update its shortest cost distance to z as three. Similarly, x
1:22 will receive a distance vector from z, five two zero, but of course, when it uses
1:28 the Bellman-Ford equation to update its distances, again the distance between z and x will be
1:34 updated from five to three. We can repeat this exercise at other nodes, as they receive distance
1:40 vectors from other nodes in the topology. And quickly, every node in the network has a complete routing table. Now
1:46 when costs decrease, the network converges quickly but one problem is that when failures occurs, bad news can actually travel slowly.