A computing power distribution scheduling method and system for an edge computing platform

Abstract

The invention provides an edge computing platform computing power distribution scheduling method and an edge computing platform computing power distribution scheduling system. The edge computing platform computing power distribution scheduling method comprises the following steps of: S1, sampling a service computing power consumption peak value; S2, evaluating a service computing power demand level according to the sampled computing power consumption peak value; and S3, performing calculation power distribution scheduling according to the calculation power demand level. According to the edge computing platform computing power distribution scheduling method, a service level factor is brought into the computing power demand evaluation, the high-level service computing power demand is preferentially met, and meanwhile, low-level service interruption caused by exclusive occupation of computing power resources by high-level services is avoided, so that the low-level services occupy certaincomputing power resources, and the basic service response of the low-level services is met; and on the basis, the requirement of short-period computing power demand evaluation can be met, and the precision of computing power demand evaluation can be effectively improved.

Description

technical field [0001] The present invention relates to a computing power distribution and scheduling method, in particular to a computing power distribution and scheduling method for an edge computing platform, and to an edge computing platform computing power distribution and scheduling system using the computing power distribution and scheduling method for the edge computing platform. Background technique [0002] The computing power scheduling of existing edge computing platforms is generally realized through two types of algorithms. One type of algorithm is based on two-factor estimation of business model and user scale, and the other type is to bind terminal devices and edge computing devices. [0003] The technical shortcomings of these two types of computing power scheduling algorithms for existing edge computing platforms include: First, the business model has a relatively stable demand for computing power during the project life cycle, and the estimation of computin...

AI Technical Summary

Problems solved by technology

[0003] The technical shortcomings of these two types of computing power scheduling algorithms for existing edge computing platforms include: First, the business model has a relatively stable demand for computing power during the project life cycle, and the estimation of computing power demand can be treated as a constant k. The business model k has different values, such as storage, encryption and decryption, image recognition, and VR video rendering, etc., and the operation is cumbersome; second, the user scale is within the project life cycle, and the growth trend fluctuates little, and the long-term estimation of computing power demand The impact is relatively large, such as the estimation cycle of months and years, but it cannot meet the estimation requirements of short-term computing power requirements, such as the requirements of the estimation cycle of hours, days, and weeks; third, the impact of computing power in the production environment There are many demand factors, such as user behavior preferences, scale of business promotion activities, holidays, hot events, weather, traffic and network failures, etc. The existing computing power demand estimation algorithm can no longer meet the actual computing power scheduling requirements; In the case of binding terminal devices and edge computing devices, computing power cannot be flexibly allocated and scheduled, resulting in insufficient satisfaction of high-level business computing power requirements, and insufficient low-level business computing power requirements resulting in waste of computing power

Images