<![CDATA[Software-Defined Networking (SDN) is a new concept on network computers. By seperating the data plane and the control plane, logics and algorithms for managing networks are placed on a controller. Currently, an algorithm that is being used for finding a shortest path on SDN networks is Diijkstra's algorithm. However, Dijkstra's algorithm has its downside, that is doesn't count bandwidth overhead. As a result, network congestion will occur if the packet size that is being send, is bigger than the available bandwidth. Therefore, the modification of Dijkstra's algorithm is needed. Dijkstra's algorithm is modified with available bandwidth as its cost. When searching a path, the algorithm will find a path which has a bigger avalaible bandwidth. There are three tests on topology A and B, that is finding a path 1 test, finding a path 2 test, and time convergence test. Finding a path 1 test is conducted when there is no traffic and changes of bandwidth on networks. Finding a path 2 test is conducted when there is traffic and changes of bandwidth on networks. Time convergence test is conducted by counting how much time that the algorithm needs to find a path 1 and 2 on each topology. When finding a path 1 and 2 tests, Dijkstra's algorithm succeed for finding a path based on bandwidth. For time convergence test, the algorithm works better on topology A when finding a path 1 and 2. The average time is 8.2 seconds and 13.1 seconds. However on topology B, finding a path 1 and 2 takes an average time about 93 seconds and 114.8 seconds.]]>
Copyrights © 2020
