Workflow group scheduling method in multi-cloud environment and based on embodiment self-adaptive distribution and integration

Abstract

The invention discloses a workflow group scheduling method in multi-cloud environment and based on embodiment self-adaptive distribution and integration. The method includes that workflow tasks are compressed by means of preprocessing, so that execution time of an algorithm is reduced; a task deadline dynamic dividing method based on embodiment execution performance is designed, so that utilization rate of execution embodiments is increased from the local level of single workflow; corresponding execution embodiment resources are dynamically distributed and integrated on the basis of performance needs of a current workflow group, so that the execution embodiment utilization rate is increased globally, and cost expenditure is reduced; task sets are dynamically scheduled onto corresponding embodiments for execution according to a principle of prioritizing latest deadline, so that each task is ensured to be executed completely before a corresponding sub deadline. The method is utilized toconduct in-depth study on optimization scheduling and embodiment self-adaptive distribution and integration of workflow groups with deadlines locally and globally, so that the performance needs of theworkflow groups are met while resource utilization rate is increased, and execution cost expenditure is lowered.

Description

technical field [0001] The invention relates to the field of parallel and distributed high-performance computing, in particular to a workflow group scheduling method based on instance self-adaptive allocation and integration in a multi-cloud environment. Background technique [0002] In the cloud environment, the instance self-adaptive allocation and integrated scheduling mechanism needs to balance the performance requirements of the workflow group and the system cost expenditure. In the face of unpredictable workflow groups, the scheduling mechanism needs to ensure that the minimum amount of instance resources that the workflow group completes before the corresponding deadline is provided, and the tasks of the workflow are scheduled to be executed on the corresponding instance resources. In case of oversupply, redundant instances are shut down in time to reduce costs. In a multi-cloud environment, each cloud service provider provides many different types of instance resour...

AI Technical Summary

Problems solved by technology

Malawski et al. designed a static and dynamic scheduling algorithm for workflow groups based on double constraints of budget and deadline, which took into account the uncertainty of task execution time in the workflow, the delay of virtual machine startup and other factors, and utilized the key workflow The admission technology guarantees the completion rate of the workflow group under the premise of double constraints. This work has a certain reference effect on the consideration of task execution time and virtual machine startup delay factors in the workflow group scheduling process of the present invention, but it only considers a virtual machine instance types, and did not discuss multi-type instances in a multi-cloud environment

Images