Research On Adaptive Load Balancing Resource Allocation Algorithm Based On Ant Colony Algorithm In SDM-EONs

Elastic optical network has fine spectrum distribution granularity,and can adaptively allocate spectrum resources according to the bandwidth requirements of the service.Therefore,it is regarded as an effective solution to cope with the increasing traffic of backbone n etwork.On the other hand,the channel capacity of single-mode fiber is growing slowly because the transmission capacity of single-mode fiber is close to the nonlinear Shannon limit of existing single-mode fiber.As a new transmission technology,space division multiplexing technology supported by multi-core fiber or multimode fiber has been introduced into the research of elastic optical network,which is used to cope with the explosive growth of broadband applications such as cloud services.The spatial distribution problem of space division multiplexing elastic optical network is extended to spatial dimension.The increasing complexity of resource allocation algorithm and the aggravation of spectrum debris generation pose new challenges for the research of network resource allocation.Among them,the resource allocation of space division multiplexing elastic optical network based on multi-core fiber is the most widely concerned,namely routing,spectrum and core assignment(RSCA).In addition,the traffic in the space division multiplexing elastic optical network also requires higher transmission performance.Therefore,the research content of this paper is to optimize the design and performance analysis of load balancing adaptive,fragment suppression and low crosstalk resource allocation algorithm in space division multiplexing elastic optical network.The main research work of this paper is as follows:(1)In order to solve the problem of load imbalance in the allocation of space division multiplexing elastic optical network resources,a load balanced adaptive routing allocation(LARA)algorithm is proposed based on the ant colony algorithm design.The algorithm includes adaptive load balancing strategy based on ant colony algorithm and dynamic static combination routing allocation strategy to monitor real-time load state of links.In the process of routing planning,the paper optimizes the uneven load distribution caused by traditional routing algorithm,considering the physical transmission cost of links and the real-time resource occupancy of each link in the network.The simulation results show that Lara algorithm has been optimized greatly in terns of network congestion rate parameters compared with the classical routing algorithms such as shortest circuit and shortest hops,and has good load balancing performance.Lara algorithm takes the utilization rate of real-time link spectrum resources as the core evaluation standard.The algorithm verifies the effect of load balancing by means of average load and variance of link load.It ensures the load balancing adaptation at routing distribution level by adjusting the low complexity,and then optimizes the network transmission blocking rate.The simulation results show that the load balancing effect of Lara algorithm is better,and the Lara algorithm achieves a 12.5%-41.6%reduction in network congestion rate compared with the traditional routing algorithm.By setting different load balancing thresholds for real-time monitoring of network status,the time complexity of the algorithm is further reduced and the performance of the algorithm is improved.The simulation results show that Lara algorithm has a good load balancing effect when the threshold is 0.9.(2)In order to solve the problem of fragment generation in the spectrum allocation of space division multiplexing elastic optical networks,the Ca-FSA(Crosstalk-aware Fragmentation Suppression Algorithm)based on fragment suppression is proposed.Due to the limitation of spectrum adjacency and spectrum continuity,the allocation of resources in different spectrum gaps will produce spectrum fragments of different locations and sizes.The algorithm can reduce the generation of spectrum debris from the source and avoid the request interrupt of defragment algorithm.Based on the design idea of fragment suppression,Ca-FSA algorithm considers the size of spectrum fragments and the duration of spectrum fragments generated by different spectrum allocation schemes,and defines the link fragmentation index(LFI)to describe the fragmentation degree of links in different situations.In the case of random arrival and random departure,the spectrum fragments in lower If links have a greater probability that the requests of subsequent arrival are reused or released earlier.Ca-FSA algorithm evaluates the fragmentation degree of link by LFI,and selects a spectrum allocation scheme with a small fragmentation degree to improve the utilization of spectrum resources and reduce the network blocking rate.The simulation results show that Ca-FSA algorithm can effectively predict spectrum fragments on links and achieve fragment suppression under the classical topology.Compared with the classical spectrum algorithms such as FF and LF,the network transmission congestion rate is reduced by 10%-40%and the link fragment is reduced by 5%-25%.(3)Combined with the spatial characteristics of MCF eons,the single fault protection strategy and crosstalk reduction technology to ensure network survivability are optimized,and a crosstalk aware single fault protection algorithm is proposed.The elastic optical network based on MCF can further expand the network capacity,but the inter core crosstalk of multi-core fiber will affect the quality of signal transmission.Based on the physical structure of multi-core optical fiber,this paper proposes a crosstalk aware single failure protection strategy.Firstly,the working core and backup core are divided according to the spatial structure of multi-core optical fiber.Based on the idea of non overlapping k-sp(k-shortest path),the backup path with lower crosstalk is selected to complete the backup of routing level,and the 100%single failure protection is realized by means of equivalent path.Next,we define the spectrum gap repeatability to describe and predict the number of spectrum gaps,and propose a spectrum layer single fault protection algorithm.Simulation results show that compared with RSCA algorithm without crosstalk sensing,the single fault algorithm with crosstalk sensing can achieve up to 17%crosstalk optimization.