Network Layer Congestion Control to Ensure Quality of Service (QoS) in Secure Battlefield Mobile Ad Hoc Networks

Abstract

To address the quality of service issues in a mobile ad hoc network (MANET) for the diverse set of applications that are carried in the network, the US ARMY CERDEC PILSNER program utilizes a Congestion Control Agent (CCA) as the focal point to facilitate QoS control in a MANET. The main functionalities of CCA include monitoring the traffic loading and queue size status of a router, and based on the measurements, CCA controls the router behavior through dynamically resetting the Weighted Fair Queuing (WFQ) and Random Early Detection (RED) parameters. In this paper, we first describe the main functionalities of CCA, including (1) Dynamically computing the RED parameters-by measuring the queue size statistics, RED parameters are dynamically adjusted based on the congestion status, such that TCP congestion window is appropriately controlled. (2) Dynamically computing the WFQ parameters-WFQ parameters are adjusted according to the measured traffic loading for the traffic classes and precedence levels. The underlying algorithm is designed to satisfy the Multi-Level Precedence and Preemption (MLPP) requirement that is unique to military applications. We investigate the performance gain achieved by CCA using OPNET simulation. We show that CCA can manage the MLPP requirements and improve the user-perceived QoS measures like throughput and call success rate.