Workflow based mixed renting method of cloud computing resources

Abstract

The invention discloses a workflow based mixed renting method of cloud computing resources. A mode that resources are rented according to requirements based on time intervals and a mode that resources are rented by bidding are included. The method mainly comprises the steps of A) offline task scheduling of a mixed renting mode; and B) online operation and dynamic adjustment of tasks. The step A) is mainly characterized in that when time of a task matches that of a virtual machine, a unit time cost priority rule is provided, the utilization rate of the rented virtual machine is improved, and a renting according to requirements mode and the bidding renting charging mode are mixed to reduce the total cost; and the step B) is characterized in that the virtual machine is rented according to requirements, a dynamic bidding price adjust strategy and a bidding failure tolerance mechanism are provided, and the system stability is improved. According to the invention, a proper sub task floating time is calculated, a virtual machine renting mode is selected, a time slot is matched, the bidding strategy and the tolerance mechanism are provided, and the operation cost of the whole workflow application is reduced greatly.

Description

technical field [0001] The invention relates to a workflow-based cloud computing resource hybrid leasing method, belonging to the technical field of cloud computing resource scheduling. Background technique [0002] Cloud computing is a large-scale distributed computing model, which provides resources (such as networks, servers, storage, software, etc.) to users in the form of services. These resources are highly scalable, scalable and available, and the biggest advantage is that for users, they can quickly reduce hardware costs and increase computing power and storage capacity, etc., and obtain extremely high computing quality at a very low cost. . Due to the above advantages, many companies and research institutions are migrating real-time analysis, scientific computing applications, etc. to cloud computing to solve data processing problems of large-scale data volumes. Therefore, how to achieve a reasonable allocation between computing or storage tasks and leased resourc...

AI Technical Summary

Problems solved by technology

Offline scheduling mainly adopts a single leasing mode, such as on-demand leasing to ensure stable resources; online scheduling considers the hybrid method of bidding leasing mode and on-demand leasing, and many studies have solved the problem of bidding resource price prediction, but due to Offline scheduling is not involved, resource utilization cannot be maximized, thus increasing the total operating cost

In the method of integrating the two, resources of the bidding type leasing mode are rarely considered

Images