Research On Adaptive Control Of Network Resource Of Service Function Chaining

In order to achieve economical,reliable,flexible and resilient network services,the concepts of NFV have emerged.The idea of Network Functions Virtualisation(NFV)can separate Network Functions from Network hardware so as to reduce the cost and achieve more flexibility in Network management and maintenance.In the scenario of NFV,many studies have focused on resource management of Service Function Chaining(SFC)with several Virtual Network Functions(VNF)on physical network.The current NFV deployment method involveshow to effectively deploy the service function chain in the static environment,without considering the change of user traffic.Therefore,the resource utilization rate is not good enough in the dynamic network environment,the performance of network can be quite poor,and the allocation of resources cannot be adjusted adaptively according to the change of the network environment.Nowadays,usual network is target for long-term objective,such as throughput and benefit.And the usage of network resource is also restricted by long-term indicators.Therefore,in this thesis,it is researched on how to efficiently manage the network resources in the dynamic environment in real time,complete the mapping of SFC,and adjust the access and transmission of data in real time.Thisthesis discusses this problem in two steps:(1)establishing a network model with multiple queues;(2)solving the transformation model with Lyapunov algorithm and proving the approximation rate.Firstly,an aproximate algorithm for calculating the optimal resource allocation in the network is presented,the optimal utilization rate of each path is calculated with the network load at each time known.Then,This thesis proposes a traffic access,routing and bandwidth allocation strategy based on the Lyapunov algorithm,which makes full use of all levels of queues in the network to cache the change of user traffic and minimize the total delay of each serviceoptimally.In addition,this thesis gives the proof that the algorithm can guarantee the stability of the queue,and proves that the result of the algorithm is only one constant away from the theoretical optimal value,which shows the performance of the algorithm in theory is excellent.Then the performance such as block rate,average queue length and average waiting time of each request are measured.The performance of the proposed resource management strategy is evaluated by numerical analysis,and the numerical results show the proposed strategy's performance.Finally,in this thesis,the SFC resource control algorithm under the multi-priority environment is also proposed,which effectively controls the packet loss rate.A large number of simulation experiments are carried out in this thesis.The simulation result shows that the performance of the proposed algorithm is effective in terms of data packet loss rate and average queue length.