An Analysis of Resilience Strategies in SDN Controllers Placement.
Software Defined Networks, Controller Placement Problem, NP-hard, Resilience, Simulation.
Software Defined Networks (SDN) emerged from the proposal to separate network management from the underlying infrastructure, allowing more efficient response to new market demands that need dynamism and high application variety. In the SDN paradigm, data and control plans are decoupled, offering the possibility to configure the network by software. However, this centralized management characteristic brings great challenges, including the Controller Placement Problem, which impacts SDN efficiency directly. This problem belongs to the NP-hard class and its optimization considers metrics such as costs, control capacity, inter device delays, resilience, among others. This project aimed to contribute with a quantitative analysis of resilience strategies in the controller placement problem. Therefore, an Integer Linear Programming model was proposed that considers several resilience constraints and the use of simulation techniques in networks to analyze different scenarios. The experimental results allowed us to understand the behavior of the network in the face of different placements of the controllers and variation in their number. The tradeoff analysis showed that the more robust scenarios imply greater investment, but it is possible to guarantee a good performance in more restrictive cost scenarios, if the controllers are properly allocated.