Research On Distributed Resource Allocation In Fog Computing And 5G Fusion Network

Network slicing has been regarded as one of the key drivers for 5G to support multiple services and application scenarios.This paper uses the spectrum resources provided by the communication network and the computational resources of the coexisting fog computing network to study the cross-system distributed network slicing.Quickly and accurately allocating resources between communication and computing networks is a challenge for network slicing across system resources.Therefore,to solve the above problems,this paper first proposes a novel distributed framework based on a new control plane entity,Forchestrator.In this framework,F-orchestrator can be deployed between tenants,physical infrastructure and service providers,to coordinate and control bandwidth and computing resources.Next,this paper improves the traditional Alternating Direction Method of Multipliers(ADMM),and proposes a synchronous ADMM algorithm based on traditional ADMM(Alternating Direction Method of Multipliers with Partial Variable Splitting,Dist ADMM-PVS).The synchronous ADMM optimization algorithm solves the challenges of network slicing across system resources well,but it has the problem of idle waiting time during iteration.Therefore,in view of the shortcomings of the synchronous ADMM optimization algorithm,this paper further improves Dist ADMM-PVS,and proposes an asynchronous ADMM optimization algorithm(Distributed Asynchronized ADMM with Partial Variable Splitting,Async ADMM).Finally,this paper proves that the proposed algorithm can minimize the service response time of the entire network,and at the same time ensure that the type of services supported by the network achieves satisfactory latency performance.Finally,simulation results show that the synchronous and asynchronous ADMM optimization algorithms proposed in this paper can converge to the minimum value in the first few iterations,which has better convergence performance than existing algorithms.At the same time,compared with the network slicing architecture that only slices bandwidth resources,the cross-system joint network slicing architecture of bandwidth and computational resources proposed in this paper can reduce the overall network delay by an average of about 13.2%.Compared with the network slicing architecture that only slices computational resources,the joint network slicing architecture proposed in this paper can reduce the overall latency of the network by about 12.5% on average.