Self-adaptive composite workflow method used in cloud computing environment

Abstract

The invention relates to a self-adaptive composite workflow method used in a cloud computing environment. The method includes the following steps that a plan workflow and a rule construction class of a composite service object are established; an initial operation instance is converted into an initial computing unit through a rule constructor, wherein the computing unit comprises an actual computing instance, a return state function and a combination state function; the initial computing unit is operated, and the output end of the initial computing unit is provided with a new rule constructor and a state information sub-computing unit; context information in the sub-computing unit is updated according to the change of the computing environment after communication synchronization with a log library; the sub-calculation unit executes the computing instance according to the rule constructor and state information, and after execution is completed, the sub-computing unit generates a next-generation computing unit. The method mainly has the advantages that the operation robustness of the workflow is improved, the workload of a user is reduced, the method can be operated in multiple computing resources at the same time, and the utilization rate of the computing resources is increased.

Description

technical field [0001] The invention relates to the field of high-performance computing, in particular to an adaptive combined workflow method in a cloud computing environment. Background technique [0002] With the rapid development of technologies such as the Internet and cloud computing, more and more computing resources are gathered on the network. Using these widely distributed computing resources to construct and execute scientific computing workflows has become a hot spot in the current academic and industrial circles. question. In practical applications, the division of labor in general business services is becoming more and more detailed, and a complex task often requires multiple small tasks to be completed in a certain order. These tasks may run on different system platforms. These computing platforms have different self-consistent strategies such as access control and transaction processing, and the status and nature of their services are constantly changing. At...

AI Technical Summary

Problems solved by technology

However, for some complex workflows, such as the need to use multiple computing modules to complete a computing task, the scheduling capabilities of the existing computing system cannot be fully utilized, and users often need to modify the program script to achieve

For example, slurm will allocate a node for each job separately during scheduling, that is, multiple jobs cannot be allocated to one node, which leads to a lot of computational waste when running small jobs with fewer cores

Images