| With the advancement of communication technology and the development of 5G networks,diverse network service applications are emerging.The development of multimedia services such as online conferences and webcasting has caused a surge in data traffic and increased the burden on the network.Multicast communication can save network bandwidth resources through multiplexing,realize efficient transmission of multimedia data,and reduce server load pressure.At the same time,network function virtualization technology can effectively reduce the cost of network equipment and improve network operation efficiency by decoupling network functions from underlying hardware devices,which is the main trend of future network development.Under the network function virtualization architecture,multicast services are implemented by deploying multicast service function chains.However,the process of determining multicast routing under the condition of satisfying business requirements and resource constraints involves the construction of multicast trees,which is one of the main challenges in the multicast service function chain deployment problem.In addition,there is a phenomenon of dynamic user access in multicast services.At this time,the topology of the multicast service function chain changes,which may lead to a suboptimal state of the overall solution;on the other hand,multicast services such as multi-person collaboration have high requirements for delay.When deploying a multicast service function chain,it is necessary to consider both end-to-end delay and delay consistency constraints between users.Therefore,under the premise of considering the network topology and user service quality,how to adjust the multicast service function chain deployment scheme according to user behavior and how to deploy the multicast service function chain while ensuring consistent user delay is the problem to be solved and the challenge faced by the current multicast service function chain deployment.Aiming at the above-mentioned problems and challenges,this thesis studies the dynamic multicast service function chain orchestration technology based on network function virtualization,The main research work is as follows:Firstly,for the scenario where users dynamically access multicast services,this thesis studies the deployment and adjustment of dynamic multicast service function chains based on user behavior.With the goal of minimizing the total network cost,comprehensively considering the migration of virtual network functions,link cost and migration cost,limited and changing underlying network resources,etc.,establish a user behavioraware dynamic multicast service chain deployment-adjustment model.Then,a heuristic algorithm is designed for this model,and a simulation experiment is designed.The experiment changes the network topology,the length of the multicast service function chain,the total number and proportion of multicast service function chains in the network,etc.,according to the network cost and actual running time that the proposed al gorithm,variant algorithm and the solver complete the deployment and adjustment,evaluate the algorithm performance.The simulation results show that the proposed algorithm is more efficient in deploying and adjusting,and can achieve the performance of deploying and adjusting at the lowest cost of 1.067 times in polynomial time in a scale-free network of 48 nodes.Secondly,aiming at the multicast service scenario requiring consistent delay between users,this thesis studies the deployment of multi-source multicast service function chains for consistent delay based on multisource underlying physical network.With the goal of minimizing the total network cost,comprehensively considering end-to-end delay constraints and delay consistency constraints,limited physical nodes and link bandwidth resources,etc.,establish a multi-source multicast service chain deployment model.Then,a heuristic algorithm is designed for this model and the simulation experiment is designed.The experiment verifies the feasibility of the algorithm according to the total network cost,actual running time and acceptance rate consumed by the deployment of the proposed algorithm and the variant algorithm under different conditions.The experimental results show that the proposed algorithm can achieve a higher acceptance rate and realize effective deployment at almost the same total network cost as the above problems.In summary,based on the network function virtualization architecture,this thesis focuses on the deployment and adjustment of user behavioraware dynamic multicast service function chains,and the deployment of multi-source multicast service function chains with consistent delay.Based on the actual multicast service scenario requirements,raises new issues and provides new solutions for the deployment of multicast service function chains. |
