Research On Technologies Of Resource Management For Cloud Testing Of Embedded Systems

With the wide application of embedded systems in the fields of intelligent driving,aerospace,medical devices and communication,their quality and reliability are gradually gaining attention.Embedded system testing,as an important technology for quality assurance,has been developed by leaps and bounds.However,the limited resources of embedded devices cannot meet the diverse testing requirements,leading to the problems of low test task execution efficiency and high testing cost.Resource management of embedded system testing in cloud environment is an effective solution,and thus how to effectively manage cloud resources in testing becomes the key.To address this issue,this thesis investigates the scheduling and migration of resources in the cloud environment and the resource management system of the cloud test platform for embedded systems,and accomplishes the following.(1)In order to reduce test task execution time and test cost,this paper proposes a cloud resource scheduling algorithm based on an improved ant colony algorithm? First,considering the influence of test task dependencies on task execution time and cost,the scheme analyzes and adopts a directed acyclic graph model to deal with test task dependencies;then introduces time and cost constraint functions in the update mode of pheromone to make the ants improve the search efficiency of the algorithm while taking into account time and cost;finally,introduces load balancing factors in the calculation of transfer probability to solve the data center Finally,a load balancing factor is introduced in the calculation of transfer probability to solve the problem of unbalanced load of virtual machines in data centers.The results of the experimental comparison through Cloud Sim simulation platform show that the performance of the algorithm proposed in this paper is better than the benchmark algorithm in three aspects: time,cost and load balancing.(2)In order to reduce the energy consumption of the cloud platform and meet the requirements of the service level agreement,this paper proposes a virtual machine migration model based on an improved ant colony algorithm.First,in order to solve the problem of large solution space of the ant colony algorithm,the model reduces the solution space searched by ants and the number of migrations by improving the VM migration strategy.Simulation experiments are conducted through Cloud Sim,and the algorithm outperforms the benchmark algorithm in three aspects: reducing the number of migrations,reducing the energy consumption of the platform and the service level agreement violation rate.(3)In order to enable better interaction between administrators and testers for resource management of the cloud platform,the proposed resource scheduling algorithm and VM migration model are applied in the self-built Open Stack cluster platform to design and implement the cloud resource management software module of the cloud platform,and test experiments are conducted to verify that the proposed method enables the test platform to manage cloud resources more efficiently.