Research On QoS Routing Extensions And Algorithms Of OSPF

Traditional Internet only provides"Best-Effort"services for data transmission. Facing to the increasingly growing up of multimedia applications, existing routing mechanisms gradually can't satisfy new demands. How to extend routing protocols in order to provide effective QoSR is becoming a considerable and investigable problem for modern Internet.This dissertation studies the working mechanisms of OSPF Protocol and implements the QoS routing extensions based on Genetic-Ant Combination Algorithm of OSPF.The dissertation gives a survey about the current researching situation of QoS routing mechanisms, bats around the existing QoSR algorithms, discusses the main problems of them and applies the Genetic-Ant Combination Algorithm to solve multi-constrained QoSR. The Combination Algorithm is based on basic Genetic Algorithm and Ant Algorithm,overcomes respective disfigurement of them. Synoptically, it initializes the pheromone value of Ant Algorithm through the optimization results of Genetic Algorithm and accordingly figures out the multi-constrained QoSR problems by iterative loop to get the best result.In order to achieve the QoSR extension of OSPF, the dissertation detailedly probe into the working process and routing algorithms of OSPF. As a typical link-state protocol, OSPF is based on Dijkstra Algorithm, which requests a certain fixed link state to compute the shortest path. This brings the result that the current OSPF Protocol can't support multi-constrained QoSR mechanism. And the mission of this dissertation is to implement OSPF-QoSR.The dissertation puts forward the idiographic implementary scheme of OSPF-QoSR. It furthest implements multi-constrained QoSR on the basis of Genetic-Ant Combination Algorithm by improving the current OSPF message format and work mechanism with minimal modification. This routing algorithm is operating in an Autonomous System, makes use of distributing routing strategy, adopts pre-computation method, extends OSPF message format to append network resource information inclusively, ameliorates LSA transmission mechanism and utilizes combination algorithm to choose the best routing path.Lastly, an OSPF network supporting QoS is constructed by OPNET Modeler to simulate the OSPF-QoSR mechanism based on the Combination Algorithm which is presented in the dissertation. The Genetic-Ant Combination Algorithm is compared in some network capabilities with the Extended Bellman-Ford Algorithm which is recommended by RFC2676. The simulation results show that Combination Algorithm is feasible and superior in contrast with the RFC-recommended algorithm. All these offer some new thoughts of multi-constrained QoSR research in large-scale OSPF networks and indicate further research orientations and emphases.