A Machine Learning Approach to Implementation of Link Aggregation Control Protocol over Software Defined Networking

Abstract

Software Defined Networking (SDN) is a complete and directly programmable network model which splits the control plane to the network data plane. Link Aggregation (LAG) is the grouping of multiple links into a single aggregated logical link with a higher bandwidth of aggregated data. This research sets out the implementation of the Link Aggregation Control Protocol (LACP) on SDN using Mininet Emulator. OpenvSwitch (OVS) acts as a transfer function, while RYU acts as an OpenFlow controller. Mininet Emulator, which is installed on Ubuntu Virtual Machine (VM) for LACP implementation in SDN. The study indicates that the speed of data communication has improved using LACP. This work also addressed that LACP provides inherent automatic redundancy that dynamically redirected to flow across the remaining links while one of the multiple links used in the aggregated groups fail or disabled. Additionally, Machine Learning (ML) approaches are also used to predict bandwidth based on statistical analysis of the data set. The Internet Service Provider (ISP) can gain more advantages to forecast bandwidth and serve customers.