ACCPndn: Adaptive Congestion Control Protocol in Named Data Networking by learning capacities using optimized Time-Lagged Feedforward Neural Network

Named Data Networking (NDN) is a promising network architecture being considered as a possible replacement for the current IP-based Internet infrastructure. However, NDN is subject to congestion when the number of data packets that reach one or various routers in a certain period of time is so high than its queue gets overflowed. To address this problem many congestion control protocols have been proposed in the literature which, however, they are highly sensitive to their control parameters as well as unable to predict congestion traffic well enough in advance. This paper develops an Adaptive Congestion Control Protocol in NDN (ACCPndn) by learning capacities in two phases to control congestion traffics before they start impacting the network performance. In the first phase - adaptive training - we propose a Time-Lagged Feedforward Network (TLFN) optimized by hybridization of particle swarm optimization and genetic algorithm to predict the source of congestion together with the amount of congestion. In the second phase -fuzzy avoidance- we employ a non-linear fuzzy logic-based control system to make a proactive decision based on the outcomes of first phase in each router per interface to control and/or prevent packet drop well enough in advance. Extensive simulations and results show that ACCPndn sufficiently satisfies the applied performance metrics and outperforms two previous proposals such as NACK and HoBHIS in terms of the minimal packet drop and high-utilization (retrying alternative paths) in bottleneck links to mitigate congestion traffics.