MILPFlow: A toolset for integration of computational modelling and deployment of data paths for SDN

Software Defined Networking is one of the most promising approaches to the deployment of future network infrastructures. The most of the Internet service providers have to deal with a number of configurations to a crescent amount of network devices. SDN is a paradigm that proposes the separation of data forwarding plane from the data control plane. OpenFlow is an standard protocol used in SDN for establishing communication among switches and controllers. However, computational modelling for SDN is still few researched in the literature. Computational modelling is the key to describe, evaluate and analyse the most diverse computational problems before its prototyping. In this paper, we propose the integration between computational modelling and the deployment of data paths for SDN. We develop the toolset called MILPFlow to generate computational models of data centers to solve routing problems and to establish data paths between servers according to the solutions of these models. Additionally, data paths are set before data flows are sent through the network. As a consequence, MILPFlow contributes to reduce the overhead to discover network routes among hosts of data centers. We evaluate the effectiveness of our methodology by using Mininet through a set of experiments.