Resource prediction and dispatching method based on local storage

Abstract

The invention provides a resource prediction and dispatching method based on local storage. The method comprises the following steps of obtaining resources of a cloud server stored in the local position; performing demand prediction on the resources; performing off line screening according to the demand prediction result so as to obtain a server to be off line; sequencing living examples in the server to be off line; selecting an immigration server for the sequenced living examples; migrating the living examples in the server to be off line into the corresponding immigration servers. The method provided by the embodiment can meet the smallest resource supply of the user service quality demand, so that the operation loss of the cloud calculation platform can be effectively reduced; the resource cost-utility ratio is improved.

Description

technical field [0001] The invention relates to the technical field of cloud computing, in particular to a resource prediction and scheduling method based on a local storage cloud server. Background technique [0002] At present, cloud server has become an important infrastructure for Internet technology and business support. It enables users to obtain computing, storage, network and other resources from a shared and configurable resource pool on demand through the network, which greatly reduces user The investment in infrastructure resources allows you to quickly deploy your own applications and run them online. At present, the research on resource scheduling and utility ratio of cloud servers has become one of the current hot spots in cloud computing technology and academics, including user resource utilization, resource scheduling process, service response time and service queue length. Various invalid losses during the operation of the computing platform to achieve "gre...

AI Technical Summary

Problems solved by technology

For example, at the current moment, the load of a certain host is too high, and one or several of its utilization indicators exceed the threshold defined by the implementation. At the next moment, the load of the host decreases, and various indicators of its utilization are below the threshold. The current high load of the host machine is only a short-term fluctuation, and there is no need to migrate the virtual machine. However, the traditional load balancing algorithm lacks a prediction mechanism for the future load trend of the host machine, which can easily cause unnecessary migration of the virtual machine.

Images