Adaptive quality of service resource management in IP networks

Abstract

Summary form only given. Different quality of service (QoS) requirements have intensified research in dynamic and adaptive communication protocols in order to efficiently exploit high speed network technologies (e.g., gigabit and terabit networks). To support guaranteed application services, IP routers/switches need to support different protocol functions or mechanisms. The existing implementation of standard protocols do not efficiently exploit the resources offered by high-speed networks and consequently, cannot provide network applications with the required high throughput and low latency communication services. We address this limitation and present a framework to provide an adaptive resource management scheme based on the QoS requirements of networked applications in an IP network. The framework categorizes QoS mechanisms into several functions such as classification, flow control, scheduling and traffic policing. The framework provides a multitude of options to implement each of the QoS functions. We also present a real-time resource allocation algorithm (RTRM) that allocates resources to applications based on their quality of service requirements and the current state of the IP router/switch. In this algorithm, the bandwidth and memory utilization are measured and the required resources are computed in order to provide the maximum available resources to the traffic class with highest priority. The results of the experiments conducted reveal that the implementation of the RTRM algorithm guarantees the QoS requirements of applications when compared to traditional nonQoS resource allocation algorithms.