Video monitoring cloud resource prediction method and system

Abstract

The invention provides a video monitoring cloud resource prediction method and system. The method comprises the following steps: a user request for video monitoring service is obtained; according to the request, and an initial resource distribution amount is determined, a current cloud intelligent video processing unit CIVU corresponding to the initial resource distribution amount is built; after current CIVU operation preset time, a first usage amount, in data corresponding to the current CIVU in a database, of video monitoring service resource in a first preset period before the current operation time is obtained; other usage mounts in a resource pool are traversed, and a first historical usage amount of which similarity to the first usage mount is greater than a preset threshold value is searched; according to the first historical usage amount, a variation quantity, in data corresponding to the current CIVU operation preset time in the database, of a second historical usage amount in an adjacent preset time period behind the second preset period is determined; according to the variation quantity and the first usage amount, a final resource demand amount of the user-requested video monitoring service resource in the CIVU in the adjacent preset time period after the first preset period can be predicted.

Description

Technical field [0001] The invention relates to the field of cloud platform resource prediction, in particular to a method and system for predicting cloud resources of video surveillance. Background technique [0002] In recent years, with the development of urban and public security construction, the application of video surveillance has become more and more extensive, especially intelligent video surveillance systems have been widely used. At present, in order to meet the demand for massive video surveillance services of today's intelligent video surveillance systems, a large number of video surveillance services have begun to be deployed and run on the cloud platform, forming a brand new VSaaS (Video Surveillance as a Service, video surveillance as a service) Cloud computing mode, namely video surveillance cloud. The video surveillance cloud service platform abstracts scattered heterogeneous resources into a shared resource pool through virtualization technology, which can pr...

AI Technical Summary

Problems solved by technology

[0004] To sum up, the problem of the existing technology is: to estimate the offline test data of the video surveillance service and allocate fixed cloud resources, which will make the resources fully utilized or reduce the user experience effect; at the same time, the existing prediction model performs cloud resource prediction, computing A large amount will bring a lot of overhead to the system

Images