A Comparison of Fast and Low Overhead Distributed Priority Locks

Distributed synchronization is necessary to coordinate the diverse activities of a distributed system. Priority synchronization is needed for real time systems, or to improve the performance of critical tasks. Practical synchronization techniques require fast response and low overhead. In this paper, we present three priority synchronization algorithms that send O(log n) messages per critical section request, and use O(logn) bits of storage per processor. Two of the algorithms are based on Li and Hudak's path compression techniques, and the third algorithm uses Raymond's xed-tree structure. Since each of the algorithms have the same theoretical complexity, we make a performance comparison to determine which of the algorithms is best under diierent loads and diierent request priority distributions. We nd that when the request priority distribution is stationary, the path-compression algorithm that uses a singly-linked list is best overall, but the the xed-tree algorithm requires fewer messages when the number of processors is small and the load is high (100% or greater). When the request priority distribution is non-stationary, the xed-tree algorithm is requires the fewest messages when the load is 100% or greater. The double-link algorithm is better when the load is low (less than 100%), or if minimizing execution time overhead is more important than minimizing message overhead.