Font Size: a A A

Research On Optimization Strategy Of SDN Flow Table Space

Posted on:2022-01-22Degree:MasterType:Thesis
Country:ChinaCandidate:L TangFull Text:PDF
GTID:2518306536463794Subject:Computer Science and Technology
Abstract/Summary:PDF Full Text Request
The control plane and data plane of the traditional network architecture are highly coupled and cannot adapt to the flexible configuration and management requirements of the current network.Software-Defined Networking(SDN),as a new generation network architecture,realizes the decoupling of forwarding and control,and facilitates the dynamic configuration of the network.Under the SDN architecture,the flow table is an important component to complete flow forwarding.At present,the flow table in SDN uses TCAM as the storage medium,and the space is limited.If the flow table space is used unreasonably,it will increase the overflow probability of the flow table,reduce the hit rate of the flow table,and affect the performance of the network,therefore,the optimization of the SDN flow table space is a problem worthy of research.From the perspective of the overall utilization of flow table space,this thesis studies the allocation and elimination mechanism of flow table space,in order to obtain the optimal flow table space utilization and improve the overall performance of the network.The main innovative work of the thesis includes:A Routing-Aware-guided Flow Table Space Balanced Allocation strategy RAFBA is proposed.The utilization status of the flow table space of the switches in the SDN network is judged through the Gini coefficient combined with the average utilization rate,and an improved ant colony algorithm IACA is used to calculate the flow path with different balance parameters.While ensuring the optimization of the data transmission,take into account the balanced utilization of the flow table space.A Flow Entry Elimination mechanism based on Q-learning FEEQ is proposed,which further optimizes the flow table space.The flow table space is maintained by an adaptive flow entry idle timeout elimination algorithm,and the Q-Learning reinforcement learning algorithm is introduced.The flow entry activity level is calculated based on the number of matching packets and the duration time of the flow entry.This dynamically adjusts the idle timeout of the flow entry.A simulation experiment is carried out on the proposed strategy and mechanism on the Mininet platform.The RAFBA strategy has verified its performance advantages in indicators such as network throughput and network delay.For the FEEQ mechanism,experiments show that it has better performance in indicators such as the hit rate of the flow table and the number of overflow entries.Finally,by combining the optimization of the flow table space allocation and the flow entry elimination mechanism,the overall performance of the SDN network can be improved.
Keywords/Search Tags:Software-Defined Networking, Flow Table Space Optimization, Routing Strategy, Idle Timeout
PDF Full Text Request
Related items