Research On Resource Allocation And Scheduling In Cloud Environment

As the scale of testing continues to increase and the test scenarios become more complex,how to quickly complete the planning of the test environment and rationally allocate various test resources according to the test requirements has become one of the challenges of cloud testing.This topic has carried out environmental planning technology research for test requirements.First,analyze the resources in the test and design scripts to formally describe the test resources.Then,analyze collaborative facility filtering technology and propose Test Facility Recommendation based on Collaborative Filtering(TFR-CF).The algorithm fully considers the topology relationship of the nodes and the user's historical habits.According to this,it recommends suitable planning schemes for users,and supports online learning based on feedback.The knowledge base is updated in time according to user feedback,which greatly saves the time for manually selecting resources.Cost and labor costs provide strong support for the test environment.Another challenge facing cloud testing is to test load balancing and fault tolerance techniques in execution.Aiming at these two problems,this paper proposes a virtual machine resource scheduling technology based on genetic algorithm.Its scheduling strategy fully considers the influence of historical factors,and combines the current state for virtual machine deployment,thus reducing the migration due to virtual machines.The test data disturbance caused by the test system improves the fault tolerance of the test system,enabling the system to provide test data with sufficient quantity and precision when some resources are abnormal.In addition,this topic also carried out research on the application of Spark Streaming technology to improve the fault tolerance of test systems.The use of Spark Streaming technology's pre-write log mechanism greatly facilitates business site recovery and data maintenance.Finally,The research content of this paper has been integrated into the experimental planning software and experimental management software of a project,which helps the planning software to improve the efficiency of the test environment construction,and helps the management software to improve the virtual machine scheduling effect and the fault tolerance of test execution.