Research On Multiple Obejective Constraint-based Routing Algorithm In Support Of Failure Recovery

With the fast development of the Internet, the network applications require higher quality ofservice, which does not only satisfy several constraints in the process of looking for path, butalso needs to enhance the utilization of network resources. However, traditional routingalgorithms can not satisfy requirements of different users effectively. Constraint-based routing(CBR) provides a new pattern for implementing better QoS mechanism.This paper mainly researches Constraint-based routing and combines with National 863project,"fast self-recovery routing protocol and test system". Through researching existingconstraint-based routing algorithms, this paper analyzes the problems of CBR algorithms. Inorder to solve these problems, this paper proposes a new constraint-based routingalgorithm—Path Resource Optimal algorithm in support of Multi-mode Protection switching(MPSPRO). This algorithm can dynamically choose optimal path according to business stylebefore satisfies multiple constraints, which can better implement optimal use of networkresources. When the network appears failures, this paper can choose the suitable protectionswitching for the working path, which enhances the flexibility and effectiveness of failurerecovery. The main content of this paper includes following aspects:This paper proposes a network model based on packet scheduling. This model combinespacket scheduling of the node with the physical link, and QoS parameters is the function ofavailable bandwidth,traffic shaping,packet scheduling and so on.This model better reflects thebasic capability of the network and is the premise of enhancing the performance of CBRalgorithm.This paper proposes a new constraint-based routing algorithm—Multiple Objective PathResource Optimal Algorithm (MOPRO) through considering the network model based on packetscheduling and multiple objective programming. This algorithm normalizes QoS parameters andchooses the corresponding normalized formulas to compute paths according to business style.MOPRO can allocate buffer and bandwidth during the process of searching paths, which canimplicitly implement load balancing and enhance the utilization of network resources during theprocess of optimizing multiple QoS parameters and searching the better path.MOPRO algorithm is the general form of considering packet scheduling in CBR, and theimplementation relies on the specific packet scheduling algorithm. This paper proposes MultipleObjective Path Resource Optimal Algorithm based on WF2Q packet scheduling style, analyzesthe computation complexityand simulates this algorithm.In order to enhance the flexibility and effectiveness of failure recovery, this paper proposesMulti-mode protection switching failure recovery method, which chooses the suitable protectionswitching style for the working path through QoS parameters and business styles. Combining themulti-mode protection switching failure recovery method with multiple objective path resourceoptimal algorithm, this paper proposes Path Resource Optimal algorithm in support ofMulti-mode protection switching (MPSPRO) , which can compute the working path and the backup path based on choosing protection switching style. This algorithm implements theeffective combination failure recovery and path searching and better guarantees the QoSperformance.This paper implement MPSPRO algorithm in the MNS mode of NS2 simulation software.The simulation results show this algorithm makes the advantage of optimally researching paths,using network resources in reason and effectively dealing with failures.