Method and device for task scheduling in a distributed system

Abstract

The embodiment of the invention provides a task scheduling method and device for a distributed system. The distributed system comprises a computing framework, computing resources used by the computing framework to execute a computing task and a cache resource pool, wherein the cache resource pool comprises affinity resources which are the computing resources firstly used by the computing framework to execute the computing task. The method comprises the following steps: receiving a resource request sent by the computing framework which has corresponding framework identifiers; judging whether the cache resource pool has the affinity resources corresponding to the framework identifiers or not; if the cache resource pool has the affinity resources corresponding to the framework identifiers, allocating the affinity resources to the computing framework for executing the computing task. The task scheduling method and the device are used for improving the operating efficiency when the computing framework executes the computing task.

Description

technical field [0001] The invention relates to the technical field of the Internet, in particular to a distributed system task scheduling method and a distributed system task scheduling device. Background technique [0002] Cloud Computing (Cloud Computing) is the growth, use and delivery model of Internet-based related services, usually involving the provision of dynamically scalable and often virtualized resources through the Internet. Cloud computing is a pay-by-use model that provides available, convenient, and on-demand network access and access to a configurable shared pool of computing resources (resources include networks, servers, storage, application software, and services) , these resources can be provisioned quickly with little administrative effort or interaction with the service provider. [0003] In the field of cloud computing, distributed task scheduling systems are very common. Usually, a physical machine or a virtual machine is used as a computing unit,...

AI Technical Summary

Problems solved by technology

However, the traditional scheduling strategy is a stateless computing resource scheduling strategy. Generally, it is known that computing nodes have computing resources, and they are randomly provided to the computing framework, which often leads to uneven distribution of computing resources.

For example, some computing frameworks may occupy computing resources for a long time, and some computing frameworks may be in a starvation state for a long time, and the task operation efficiency is low

Images