Software-Defined Networking (SDN) is an emerging architecture that is dynamic, manageable, cost-effective, and adaptable, making it ideal for the high-bandwidth, dynamic nature of today’s applications. This architecture decouples the network control and forwarding functions enabling the network control to become directly programmable and the underlying infrastructure to be abstracted for applications and network services. The OpenFlow® protocol is a foundational element for building SDN solutions.
Traffic flow prediction is an important research issue for solving the traffic congestion problem in any network. Which leads to over or underutilization of resources further leading to packet loss , jitter, severe performance degradation, and economic losses, etc. Traffic congestion can be predicted in advance by analyzing traffic flow patterns. Such prediction is possible by analyzing the real-time network data, which can be used to adapt to the needs accordingly by allocating or de-allocating resources. Also, techniques such as network function virtualization(NFV) and switch migrations (in SDNs) can be used to tackle such problems.
SDN architecture and the built-in data collection mechanisms of the OpenFlow® protocol makes it easy for us to collect data without any need for extra middle-ware. Firstly, We will understand the type of statistics that can be collected in SDNs and how this information can be used to learn from the collected data. Secondly, we will apply Random Forest (RF) and Xtream Gradient Boosting (XGB) prediction Techniques on the collected datasets to make traffic predictions. Finally, we will compare both the techniques and use the technique with the best results to develop an application which can run on top of the controller to predict such traffic congestion.
For set-up of this project Mininet network emulator is used as a to emulate the switches in the SDN network and the controller used is Ryu because of it's a wide collection of libraries, active community, frequent updates and its support for all the OpenFlow® protocols. Ryu also provides with a built-in application which is used to get different network statistics from the controller. In this project a virtual SDN will be set up using Mininet then traffic will be generated to simulate real network traffic and datasets will be generated by aggregating the statistics from the controller. The final step is to compare the RF and XGB techniques and find which technique can be used to accurately predict the future traffic load.