Privacy-Aware Switch-Controller Mapping in SDN-Based IoT Networks
Vignesh Sridharan, Kalupahana Liyanage Kushan Sudheera, Mohan Gurusamy · 2020
Software Defined Networking (SDN) based solutions have been proposed to address challenges related to IoT traffic management and security. However, the SDN architecture introduces new threats that can impact the privacy of IoT devices. A compromised controller in the network can infer sensitive information by analyzing the network traffic of IoT devices and predict future user behavior. To minimize the information availability to controllers, we propose DPAM, a Dynamic Privacy-Aware Mapping approach, which changes switch-controller mapping in a scheduled manner to prevent controllers from monitoring the same switches continuously for long durations. As a part of DPAM, we formulate an optimization problem whose solution generates a switch-controller mapping schedule that minimizes the maximum information that is available to a controller from a switch while reducing the overhead incurred due to frequent mapping changes. We obtain optimal solution to our problem and evaluate DPAM's performance by comparing it against other dynamic mapping approaches. We demonstrate in our performance study that DPAM reduces the maximum amount of information available to a controller by up to 40% as compared to the other approaches, while achieving better controller load balancing.