Task scheduling-based load balance method and system

Abstract

The invention discloses a task scheduling-based load balance method and system. A central task scheduler adopts a load balance algorithm; executors establish Deques and store received tasks in the Deques; and the thief executor captures the tasks at the head of the Deque of the Victim executor through a bisection method according to a received address of the Victim executor, and inserts the tasksinto the own Deque. Through modes of task distribution, task extraction and the like, the executor with a relatively small task quantity is scheduled to execute the tasks of the executor with a maximum task quantity, so that dynamic distribution and migration of the tasks are realized, the load imbalance phenomenon of each executor is eliminated, and the task execution becomes efficient and parallel; and each executor stores the tasks through the Deque and extracts the tasks through the bisection method, so that the problem of task execution deadlock caused by resource contention is avoided, the higher load balance degree is obtained, and the parallel resource utilization rate and the overall performance of the system are improved.

Description

technical field [0001] The invention relates to the field of cloud computing resource scheduling, in particular to a load balancing method and system based on task scheduling. Background technique [0002] Load balancing is the distribution of load among multiple computers (computer clusters), network connections, CPUs, disk drives, or other resources in order to optimize resource usage, maximize throughput, minimize response time, and avoid overloading at the same time . More and more modern enterprise information applications adopt B / S application architecture to carry the key business of the enterprise. Therefore, it is becoming increasingly important to ensure the reliable operation of these tasks. As more and more enterprises implement data centralization, the scalability, security and reliability of applications are also paid more and more attention by enterprises. [0003] In cloud computing, task scheduling is the key to affecting system performance. How to effect...

AI Technical Summary

Problems solved by technology

There are many deficiencies in the traditional task scheduling system, which cannot meet the increasing computing demands of tasks, and there are a lot of idle shared resources of some computers in the cluster, and the overload of a few computer systems is too large, which leads to the failure of each node. Shared resources cannot be fully utilized, and these problems will seriously affect the performance of the system

Images