Architecture and Algorithms for Scalable Mobile QoS

Supporting Quality of Service (QoS) is an important objective for future mobile systems, and requires resource reservation and admission control to achieve. In this paper, we introduce an admission control scheme termed Virtual Bottleneck Cell, an approach designed to scale to many users and handoffs, while simultaneously controlling hot spots. The key technique is to hierarchically control an aggregated virtual system, ensuring QoS objectives are satisfied in the underlying system without per-user resource management such as advanced reservations of bandwidth in a user's predicted future locations. We develop a simple analytical model to study the system and illustrate several key components of the approach, such as balancing the conflicting design objectives of high utilization, scalability, and ensured QoS. We formulate the problem of clustering cells into virtual system as an optimization problem and propose a heuristic adaptive clustering algorithm as a practical solution. Finally, we evaluate the scheme by developing a simple analytical model, devising an optimal off-line algorithm, and performing simulations of a two-dimensional network.