I'm doubtful about the effectiveness of the two explained routing algorithms. Link-state tries to find recursively the shortest path between a set of routers but what happens if a new router is allocated? Will all routers at global scale modify it's path by this change of performance if it is much better? Will it check if there can be more collisions because of this? And with the case of infinite time states, in the case that router receives a package that doesn't belongs to any route, it will directly drop the package or it will be sent to other router that is not on its global routing algorithm range?
And with distant vector, it will check other possible variables as the maximum packages that can be handled by the same time or the congestions? Will it try to compute the values empirically and not by theory values like the distance between the routers, the propagation speed of the medium or the maximum bandwidth?
Link-state means each router advertises to the whole network (all routers) the state of the links it is connected directly. If a new router appears, it will be noticed by all the others when receiving messages from it. All the new routes will take into account the new routes that are now possible thanks to the presence of the new found router.
The result of the process are forwarding tables. Each router will have one and it will use it to make forwarding decisions. There are no drop rules, but an unknown destination would be dropped and source informed about it (default rules will prevent that though).
Distance vector does not evaluate any constraints not taken into account into the metric used for link assessment.
I'm doubtful about the effectiveness of the two explained routing algorithms. Link-state tries to find recursively the shortest path between a set of routers but what happens if a new router is allocated? Will all routers at global scale modify it's path by this change of performance if it is much better? Will it check if there can be more collisions because of this? And with the case of infinite time states, in the case that router receives a package that doesn't belongs to any route, it will directly drop the package or it will be sent to other router that is not on its global routing algorithm range?
ReplyDeleteAnd with distant vector, it will check other possible variables as the maximum packages that can be handled by the same time or the congestions? Will it try to compute the values empirically and not by theory values like the distance between the routers, the propagation speed of the medium or the maximum bandwidth?
Link-state means each router advertises to the whole network (all routers) the state of the links it is connected directly. If a new router appears, it will be noticed by all the others when receiving messages from it. All the new routes will take into account the new routes that are now possible thanks to the presence of the new found router.
ReplyDeleteThe result of the process are forwarding tables. Each router will have one and it will use it to make forwarding decisions. There are no drop rules, but an unknown destination would be dropped and source informed about it (default rules will prevent that though).
Distance vector does not evaluate any constraints not taken into account into the metric used for link assessment.