Based on the distributed file system method and system for scheduling request io

The present invention discloses a distributed file system based on the IO request scheduling method, comprising: receiving an IO request from the client, and update the global IO request according to the process table, the global IO request process object added to the request queue to which it belongs end, and continuously determines the global IO request process object is smaller than the length REQ_MAX (L), and there are L & lt; multilevel request queue maximum number of levels, if L from the first stage of queues delete the global process object, and inserted into the end of the first L + stage 1 queue, continuously monitoring from a IO request from a client, and repeat the above steps to generate a dynamic multi-stage process request queue, extracts the current does not request queue is empty multi-stage process first All IO request process object, execute and update the status of the object information and process parameters. The method of the present invention is capable of carrying large decrease in overall system response time to client requests to a great extent, to improve the performance of the system.