Research On Reliability Redundancy Allocation Problem And Reliability Design Of Communication Networks

The softwareization of 5G networks to support services such as enhanced mobile broadband and ultra-reliable and low-latency communications has revolutionized the networking industry.Through virtualization and cloud computing,network functions virtualization helps communication networks deliver services at a lower cost and time through the service function chaining.However,due to the softwareization of network and service functions,the possibility of the failure of communication systems due to the software problems increases,which brings many new challenges to communication systems in terms of system reliability and latency reliability.Reliability optimization of communication networks is one of the important topics,and the reliability optimization of a system involves many complex and interrelated factors,one of which is to translate the reliability of the entire system into the reliability requirements of the individual subsystems.At the same time,with the commercialization and popularization of 5G communication,different application scenarios put forward different latency reliability requirements for communication services.In the face of the decline in system reliability caused by higher probability of software failure,how to use limited bandwidth resources to meet the given latency reliability requirement is also one of the important challenges faced by communication networks at present.The reliability redundancy allocation problem(RRAP)finds an optimal system structure that achieves its highest possible reliability by determining the redundancy level and reliability variables in each subsystem while satisfying resource constraints.Considering the current 5G communication scenarios,this paper hopes to construct a general reliability redundancy allocation problem suitable for modern communication networks and solve the problem more efficiently by improving the traditional meta-heuristic algorithm.In order to meet the strict requirements of 5G communication services for latency reliability,network function virtualization and software-defined networking provide some new solutions.This paper hopes to give an accurate latency reliability analysis framework and propose a system reliability design scheme that can meet the Qo S constraints of service requests and higher system reliability requirements.The main work and innovation of this paper are as follows:(1)In order to apply the reliability optimization problem to the more general system structure,this paper applies the global reliability of communication networks to the reliability redundancy allocation problem for the first time and formulates a new RRAP called global reliability RRAP(g RRAP).In order to make the optimization problem more realistic,the problem considers imperfect nodes and regards nodes and links of the system as different subsystems.To evaluate the global reliability of the general system,a recursive algorithm is proposed based on the concept of the tree in graph theory.(2)Due to the high-dimensional problems brought by global reliability evaluation,the complexity of the optimization problem is increased.Therefore,an improved particle swarm optimization algorithm with a specific particle initialization method(called PIPSO)is proposed.In PIPSO,a measure is proposed to evaluate the distance between particles,and an index is designed to quantify the uniformity of particle position.Through these methods,the proposed particle initialization method based on Centroidal Voronoi Tessellations(CVT)can make the initial position of the particle more evenly distributed in the feasible domain,thereby improving the global search ability and convergence speed of the algorithm.(3)In order to compare and illustrate the performance of the PIPSO algorithm,this paper adjusts three popular meta-heuristic algorithms to solve the g RRAP,and compares the performance gap between different algorithms.At the same time,this paper also uses the PIPSO algorithm to solve four benchmark problems and compares the results with the previous works to illustrate the advantages of the algorithm.(4)In order to analyze the latency performance of the multi-service queuing system more accurately,this paper gives a latency reliability analysis framework for multi-server queuing system based on martingale theory and gives a simulated annealing algorithm for bandwidth estimation based on this framework.Considering the software-based network environment,in order to meet the Qo S constraint of service requests and higher system reliability requirements,this paper presents a reliability design scheme of multi-server queuing system through redundant backup.