A Distributed Cache Architecture with Snooping for QoS Routing in Large Networks

Abstract

To meet the diverse quality-of-service (QoS) requirements of emerging multimedia applications, communication networks should provide end-to-end QoS guarantees. QoS routing is the first step towards this goal. The route computing overhead caused by on-demand calculation of QoS routes, especially in large networks with heavy traffic, is a concern and can cause scalability problems. This paper addresses this problem by introducing a novel distributed cache architecture. The distributed nature of the proposed cachearchitecture facilitates its deployment in large networks. To maximize the performance of the distributed cache architecture, cache snooping has been proposed to alleviate the side effects of network states fluctuations on the cached route so that the overall routing performance is significantly improved. Assuming a bandwidth-based QoS model, in performance evaluation of the proposed distributed cache architecture, we use a broad range of realistic network topologies, network traffic conditions, routing protocols, and aggregation techniques to evaluate different aspects of the proposed cache architecture under different conditions. The results confirm that the route caching is quite effective in reduction of route computing overhead. In addition, our results suggest that the cache snooping can significantly increase the overall routing performance, especially in the presence of highly inaccurate network state information.