Overview of IO Queuing Algorithm in Distributed Memory

Abstract

The distributed shared memory system provides a development road for establishment of a large-scale easy-to-use parallel computer, and the improvement of its corresponding algorithm can optimize the distributed shared memory. This paper mainly summarizes the spatial data memory system and parallel processing methods, thus providing the specific theoretical guidance for the parallel spatial data memory and processing. This paper also analyzes and improves the basic distributed shared IO algorithm of the central server algorithm, migration algorithm, read-replication algorithm and full-replication algorithm, and then focuses on discussion of the consistency model of the distributed shared memory, and analyzes the direction of using different consistency models, and points out the shortcomings of traditional methods to store and process massive spatial data, thus providing a theoretical direction for the future improvement of the memory.