Modeling and Analysis on Congestion Control for Data Transmission in Sensor Clouds

Abstract

Wireless Sensor Networks (WSNs) and Cloud computing are two emerging technologies, respectively, in the networking and computing fields. Sensor Cloud integrates WSNs and Clouds into an extended form of cloud infrastructure that manages sensors scattered throughout WSNs. Networking plays a crucial role in Cloud computing and becomes a particularly challenging issue in sensor Clouds due to resource limitation of WSNs as the underlying network of the Cloud infrastructure. Therefore, data transmission from sensors to servers in sensor Clouds must be carefully controlled to avoid network congestion. In this paper, we study Random Early Detection-Based (RED-based) congestion control for data transmission in sensor Clouds. Specifically we propose an improved RED (IRED) algorithm that meets sensor Cloud networking requirements better and develop a queueing model coupled with analysis technique to evaluate the performance of the proposed congestion control scheme. We also conduct extensive simulations to evaluate performance of IRED compared with the regular RED algorithm. Our analysis and simulation results demonstrate that the proposed IRED congestion control achieves comparable delay performance and preferable throughput performance in comparison with the standard RED for data transmission in sensor Clouds. The easily implemented control mechanism and low complexity of our proposed algorithm make it more suitable for sensor Clouds.