Research And Simulated Implementation Of Intelligent Fault-tolerant Qos Routing Mechanisms In Ngi

The development of computer network and communication technology has been changing ways of people's life and work. NGI (Next Generation Internet) which has more bandwidth resources and combines many types of wired and wireless networks, provides a more flexible, more diverse network services with higher QoS (Quality of Service). However, some characteristics of NGI, such as heterogeneity, mobility and instability, make troubles for providing reliable QoS; many mechanisms and protocols in current Internet need to be reconsidered and redesigned. Meanwhile, society depends on information much stronger now, but the current network facilities often express fragile for various reasons. Thus, how to improve the fault-tolerance of NGI becomes the problem badly in need of solution. Fault-tolerant routing mechanisms are one of the effective methods to solve the problem.In this paper, the research mainly focuses on shared-path preventive and intelligent fault-tolerant QoS routing mechanism in the case of single-SRLG (Shared Risk Link Group) failure. Four algorithms (Outlook Algorithm, Evolutionary Multi-agent Algorithm Applied to the Multi-objective Optimization Problem, PSOSA Hybrid Optimization Strategy, Tabu Fish Swarm Algorithm) are used to solve the problem of intelligent fault-tolerant QoS routing in NGI. The aims are not only to guarantee QoS of users'communication, but also the influence of link's uncertainty on the fault-tolerant QoS routing mechanisms is considered. By analyzing, comparing and referring to some existing fault-tolerant routing mechanisms, the intelligent fault-tolerant QoS routing mechanisms in NGI are divided into two steps:QoS requirements determination and fault-tolerant routing. During design, fuzzy and uncertain features of users'requirements and network parameters, dynamic features of link quality are considered. In allusion to intelligent fault-tolerant QoS routing, the objective is to maximize users's end-to-end reliability, minimize the cost of path, and maximize the value of path probability function. Partial SRLG-disjoint primary path and backup path are established to improve the success ratio of the path connection request, that is, if the primary path reliability satisfies the users's requirement then we needn't to establish a backup path, otherwise, the solution is converted to SRLG-constraint-based shared-path dynamic protection design problem. In allusion to intelligent fault-tolerant QoS multicast routing, the objective is to minimize the cost of multicast tree. The primary and the backup trees with QoS satisfied are established to realize fault-tolerant using redundant tree reconvery method. The QoS negotiation with user is necessary when the backup tree that satisfied the QoS is not found. To verify the correctness and the feasibility of the intelligent fault-tolerant QoS routing mechanisms in NGI, simulated implementations and performances analysis are made with NS2(Network Simulator 2). The simulation results have shown that our mechanisms can guarantee communication QoS in the fuzzy, uncertain and dynamic network environment compared to some existing fault-tolerant QoS routing mechanisms and the designed mechanisms are feasible and effective in case of single-SRLG failure.