The Optimization Of Container-Based Virtual Machine Scheduling Algorithm

The emergence of container technology is to solve the problem between the independence of the application and the interoperability of resources face from the operating system. It can meet the basic needs of the independence of system resources while providing efficient support. Container technology can solve the virtual machine problems in safe and resource scheduling, while providing efficient interoperability of resources and maintain procedures for the relative independence. Thus the container-based virtual machine technology is very significance.This paper analyzed the status of virtualization and through reference the core ideas of resource container, then details and analyzes the security container and the resources container, and proposed system architecture of the container virtualization technology then compared the original Hypervisor model. Finally, this paper fully endorsed the sharing of resources in the system's advantages.In addition, this paper researched in several different CPU scheduling algorithm of the Xen virtual machine, and put the most widely used and efficient scheduling algorithm Credit as the main object of study. For its response in less load conditions does not guarantee real-time, response latency was generally longer and volatile situation, Proposed a improved Credit algorithms. This paper also studied the impact of factors that Virtual machine optimize network performance, and CPU scheduling algorithm as the main entry point revealed in the importance of network performance optimization, and propose a new optimization program.In the improved scheduling algorithm for each run Guest operating system adds a virtual hardware environment ROOT state, while for each virtual machine increase of a feedback process, when the current Guest operating system in an empty load status, the process will be adopted the Guest operating system status notification Xen by the way called hypercall. ROOT state in the virtual machine has a higher priority. ROOT state is to provide a new domain mechanism to achieve low I/O and CPU latency.lt have a lot of advantages to address the high latency and Verified by experiment.This study have a certain significance on the cloud computing virtualization part of the group.