The Design And Implementation Of Mooctest Hybrid Cloud's Manage And Schedule Center Oriented To Multi-Objective Task Scheduling

The Mooctest is a public service platform that connects the educational,business and academic circles.This platform introduces the cases of enterprise requirements to the teaching,and also excavates the research hot spot to support academic research.At present,the platform has received the attention of hundreds of universities and enterprises,and has successfully supported the two national college student software testing competition.With the increasing influence of the Mooctest,some colleges,universities and enterprises for reasons of user data security,network environment and so on want to deploy a Mooctest private cloud,using resources(business needs and hardware equipment)from the Mooctest public cloud and other Mooctest private clouds at the same time.That way the optimal allocation of resources can be achieved.In order to meet this requirement,this thesis proposes a solution based on hybrid cloud architecture.This architecture combines public cloud and private cloud,and is the main mode and development direction of cloud computing in recent years.It breaks through the hardware limitations of private cloud,and makes use of the scalability of public cloud,and can acquire higher computing power at any time.After the transformation,the original platform for differentiation of public cloud and private cloud edition version,and the scheduling and management center of new connections,the private cloud in the task release way of sharing public cloud resources,but also accept other cloud's mission in my spare time to share their resources,increase the utilization rate of the whole cloud platform resources.Mooctest hybrid cloud scheduling and management center,as the core module of Mooctest hybrid cloud,provides functions such as cloud registration and management,cloud resource management,task scheduling,task result summarization and platform monitoring.Java are used in the development of Mooctest scheduling and management center under spring-boot and Hibernate framework to provide HTTP interfaces for the cloud nodes,providing functions like cloud registration and management,resources management,tasks management,results released and so on.Using the Dubbo framework to provide RPC interface for public cloud,allowing administrators to perform global monitoring and management of private cloud resources and tasks in the public cloud node.At the same time,docker and jekins are used to automate the deployment process quickly.In order to satisfy the multi-objective task scheduling requirements of users and platforms,and improve the overall efficiency of the platform,the task scheduling algorithm is also optimized.A multi-objective task scheduling algorithm is proposed,which makes the overall effect of task scheduling optimal.