BCON: Back pressure based congestion avoidance model for Named Data Networks

Abstract

Queue management in Named Data Networks (NDN) has gained a lot of attention recently. Unlike the traditional IP architecture, the transport mechanism in NDN is intricate and comprises of in-network caching of data packets at routers. Hence, the most effective avoidance of congestion can occur at the routers itself. In this paper, we make two contributions: first, we propose a back pressure based congestion avoidance model for NDN which leverages the benefits of Active Queue Management (AQM) mechanisms. Using this model, we apply the existing AQM mechanisms like Random Early Detection (RED), Adaptive RED (ARED), Controlled Delay (CoDel) and Proportional Integral controller Enhanced (PIE) in NDN. Second, we study the effectiveness of our proposed model by performing simulations using ndnSIM. Our simulation results indicate that the proposed model successfully balances the tradeoff between link utilization and Data drop rate.