Analysis of congestion control algorithm for fog computing systems with internet of things

Abstract

With rapid deployment of Internet of Things (IoT) applications which demand real-time, lowlatency services, fog computing was proposed as a promising distributed computing paradigm to support the IoT systems. Unlike the centralized cloud computing architecture, fog computing deploys network devices with various degrees of computational and storage capabilities at the edge of the network closer to the IoT devices in order to meet the realtime, low-latency service demands generated by these devices. Considering the rapid growth of IoT systems and the huge traffic generated by billions of IoT devices, there is the need to have an effective algorithm implemented at the fog gateways for congestion control. In this paper a queuing model was proposed to analyze the effectiveness of the Random Early Discard (RED) algorithm with generalized nonlinear loss function for edge gateways of fog computing architecture which supports the IoT applications. Results of the analysis have shown that in high traffic load situations, the probability of service requests’ drop increases as the index value of the generalized nonlinear loss function decreases. For light traffic load situations, the throughput of service requests increases as the index value of the generalized nonlinear loss function increases.

Keywords: RED, Delay, Congestion Control, IoT, Fog Computing.