A Queueing Theory Based Approach to QoS-Driven Adaptation for Service Discovery over MANETs

摘要

Volatility and uncertainty characterizing mobile ad hoc networks (MANETs) demand that an efficient mechanism to discover currently available services in a MANET should be provided. Consequently, many service discovery systems over MANETs have been developed. However, most of them lack theoretical modeling and analysis, as well as the leveraging of the theoretical model for optimizing system design and improving quality of service (QoS). Aiming at our service discovery system SCN4M, this paper develops an M/M+/1 model for service discovery on MANET nodes. This model is first applied to predict system behaviors such as system throughput. Then, the model is applied to find the value of a control variable. If the calculated value is set at runtime, the system will provide the optimal QoS. Furthermore, the model is used for optimizing a system's adaptive mechanism so that the system can achieve the optimal QoS no matter what kind of environmental change occurs. Experiments are also conducted to verify the model and its applications. Our approach indicates that the queueing theory model can be seamlessly combined into system design to improve system performance.