The problem of choosing routing algorithm

Routing is the process of picking the route for transmitting data in computer networks. This function is performed by the router. Router is the device where two processes are carried out. The First process is the process of handling or processing the packets and the second one fills and updates routing tables using the certain algorithms. Routing tables can be modified during the working time of network. This could be done by the routers automatically due to exchange of service information - dynamic routing, or manually by the administrator - static routing. Distance-vector routing uses routers’ tables that contain the information about all routers in the subnet. One of the problems of this algorithm is that the information about new “good” route spreads much faster than information about lost or destroyed ones. Link state routing broadcasts route information to every device in this particular network. But in this case every router sends small piece of its routing table that contains only its own links. Link state routing algorithms are much more convergent and they have lesser looping probability comparing to distance-vector routing algorithms but they require more powerful platforms because of bigger quantity of computations and higher memory requirements. Therefore link state routing is more expensive to implement than distance-vector one. The choosing of routing type is one of the most important problems for developer when designing the network. This problem leads to problem of choosing routing algorithm. Everything depends on the type of network that is needed in this particular situation. If one needs the lowest cost network without good algorithm convergence and without the best optimality his choice should be the distance-vector routing. And if the money is not the problem but low looping probability, high convergence, flexibility and optimality are the key requirements choice should turn next to the link state routing.