A quality of service (QoS) implementation of Internet cache coherence

Caching has long been employed in computer system architectures to improve performance in terms of reduced memory access times and latencies and hence, to improve throughput at the expense of additional complexity in memory organization and of managing multiple copies of shared data. In traditional large-scale computer architectures, the cache coherency schemes developed permit very large, cache-coherent non-uniform memory access (CC-NUMA) to shared memory spaces. However, these approaches fail when applied to the vastness of the Internet and the growing complexities introduced by mobility. While caching has been successfully employed in specific, limited applications in modern Internet implementations, there has been no general-purpose approach to cache coherency on the Internet. A quality-of-service (QOS) approach is ideally suited for such a general-purpose cache coherence protocol, providing strong consistency for those data items that require it while permitting weaker consistency for less critical data. A statistical analysis of the read/write behavior of typical Internet data is used to suggest a low overhead, inexpensive QOS solution to cache coherency issues on the Internet. An experimental framework studies and verifies the potential of the proposed scheme as a viable solution to cache coherence on both wired and wireless Internet applications.