Distributed-system observation tools require an efficient data structure to store and query the partial-order of execution. Such data structures typically use vector timestamps to efficiently answer precedence queries. Many current vector-timestamp algorithms either have a poor time/space complexity tradeoff or are static. This limits the scalability of such observation tools. One algorithm, centralized hierarchical cluster timestamps, has potentially a good time/space tradeoff provided that the clusters accurately capture communication locality. However, that algorithm, as described, uses pre-determined, contiguous clusters. In this paper we extend that algorithm to enable a dynamic selection of clusters. We present experimental results that demonstrate that our extension is more stable with cluster size and provides timestamps whose average size is consistently superior to the pre-determined cluster approach.
展开▼
