Research On Consistency Update Method In Software Defined Network

Flow table updates are very frequent in Software Defined Network(SDN),If the consistent update of the flow table rules cannot be guaranteed,the black hole,transmission loop and link congestion may appear in the network,thus affecting the performance of the network.How to update the flow table consistently to ensure the normal transmission of packets during the update period has been widely studied.Some existing research schemes cannot simultaneously guarantee that the update process is free of black holes,loops and link congestion,some are too complex to affect the update time,and some may consume additional switch storage resources.Aiming at the problems existing in the current research on the consistency update of flow table in SDN,the flow table update detection mechanism is designed,and the SDN controller will adopt different update strategies for the flow table update caused by different reasons.A selection algorithm for minimizing the number of data streams to be updated is proposed to avoid link congestion in the process of selecting the update path for the data stream to be updated.For each data flow to be updated,analyze whether there is a loop in the new path and the old path,and divide the different switch sets according to the flow table update operation type for the new and old paths with and without loops respectively,and take different Update strategy to ensure the consistency of updates.For the update of the flow table on the old path,the real-time flow table update method of link real-time detection is adopted to ensure the timely update of the old path to reduce the update time and ensure that the data flow packets on the old path are transmitted completely.The experiment is based on the Floodlight controller and Mininet network simulation software to build a software-defined network experiment platform.The experiments of updating a single data stream and multiple data streams in different topological environments are carried out,and compared with the rest.Under the same experimental conditions,the experimental results show that the flow table update time of this scheme is shorter,the number of data streams to be updated and the number of update rules are smaller,and the load on the controller is lower.