Research On Joint Resource Opitmization Technologies In Service-Driven Mobile Carrier Networks

With the emerging applications such as high-definition video and autonomous driving,mobile services in the 5G era exhibit high bandwidth,low latency,and high reliability.The mobile bearer network driven by the service is constantly evolving and featured by the heterogeneous integration of multiple technologies and multiple resources.At present,a key problem still exists in the mobile bearer network is how to optimize the resource configuration of the mobile bearer network to meet the needs of service bandwidth,delay,security and other aspects,while improving the economic benefits of the network.This paper focuses on the joint optimization of network resources,and aims to improve the efficiency of resource utilization in different scenarios.The optimization algorithm design and experimental verification are carried out.The main research contents and innovations include the following three aspects:(1)Propose an energy-efficient lightpath adjustment strategy for tidal services.Aiming at the problem of inefficient resource utilization caused by static network configuration under mobile tidal traffic,an energy-efficient dynamic lightpath adjustment strategy was proposed.According to the dynamics of traffic load,this strategy performs dynamic lightpath adjustment on an AWGR-based passive WDM fronthaul network architecture.The lightpath adjustment is used to aggregate the services in the low-load baseband processing unit to improve the energy efficiency.The feature of this strategy is that it adopts the energy-efficient dynamic lightpath adjustment method under AWGR-based routing rules,and considers both routing constraints and baseband processing resource constraints,and jointly optimizes the energy consumption of bandwidth and processing resources.Simulation results show that the proposed dynamic lightpath adjustment strategy can reduce network energy consumption by an average of 30%compared to static configuration.In addition,this paper also develops and verifies the control plane for the energy-efficient dynamic lightpath adjustment method.The results of this experiment show that the lightpath adjustment operation can reduce the energy consumption of baseband processing resources by 50%.(2)Propose a low-cost RAN slice deployment solution for service isolation requirements.Aiming at the isolation requirements of slicing services,a RAN slice deployment strategy based on isolation is proposed.This strategy is aimed at satisfy the service processing,bandwidth,delay and isolation requirement.The RAN function placement and traffic routing are performed through node ranking,which optimizes the network resource cost required for RAN slice deployment.The feature of this solution is that it focuses on the functional isolation and traffic isolation constraints during the slice deployment,proposes RAN slice deployment methods at different isolation levels,and explores the relationship between isolation levels and resource deployment.Simulation results show that the proposed strategy can effectively lower the deployment network resource cost,and the difference between the optimization cost obtained by the mathematical model does not exceed 10%.In addition,for different slice isolation level requirements,the results show that as the isolation level increases,the network needs to deploy up to 6 times more computing resources and 4 times more bandwidth resources.(3)Propose a RAN slicing adjustment solution for the dynamic service requirements.Aiming at the dynamic service requirements,a RAN slice adjustment strategy based on traffic prediction is proposed.According to the prediction information of mobile traffic,this strategy pre-configures resources during the slice scaling,to avoid service degradation due to lack of resources.The feature of this solution is that,compared with the non-prediction solution,this solution performs a pre-configured "dynamic"resource redundancy configuration on the slices in the network based on traffic prediction information,which improves the success possibility of slice scaling.In addition,the strategy also takes advantage of the complementarity between different traffic patterns to avoid slice migration due to slice scaing up/down.Simulation results show that,compared to the"static" resource redundancy configuration,the slicing adjustment strategy based on traffic prediction can reduce the service degradation due to scaling failure by up to 38%,and can reduce the slice migration by up to 32%.