Research And Implementation Of Improved Technologies In Virtualization Technology Based On Linux Container Resource Utilization

Virtualization technology has been widely applied in different fields.Popular VMM such as VMware,KVM,XEN,etc.They belong to the Platform Virtualization;It can be further divided into 2 types: Full Virtualization or Hardware-assisted Virtualization technology.Currently,Container Virtualization is widely used because of its low cost,good scalability.Container implements the operating system level virtualization,The new container instance simulated by the operating system kernel.It can reuse the host kernel and libraries,and without the Simulate Hardware and Interpretation Instruction.Different container processes can be isolate.The container provides a lightweight and efficient services make use of Namespace and Cgroups provided by the kernel,every container can provide the isolated environment and resources management.In addition,isolation and utilization of container resources can still be improved.In this paper,New model used to achieve Container Virtualization Isolation and Optimize Resource Utilization,the new scheme set out above are based on Linux container.This scheme used to solve problems caused by run idle or wasted resources by competition,while optimizing the container engine isolation through container grouping and filtering.Container resource limits provided by the Linux kernel Cgroup mechanism.Use each of the Cgroup subsystem parameter to manage different types of resource.Such as,through quota parameters to limit the run-time environment,Alarm and the automatic control parameters can open as Linux oom or other functions,Use statistics and monitoring parameters can provide log files.Modified LXC by add additional new function module.Then,it is capable of performing dynamic monitoring of the container state at run-time,and it can adjust real-time configuration environment,The resource allocation on demands will become reality.The running container will be managed by the regional grouping and hierarchical tree structure.And to achieve optimal performance,the scheduler will use a priority queue structures to dispatch all running container.In order to implement this model in LXC,the original platform has been modified.Specific to the structure,we add three new modules to the model.The first is the Container Group Management Module that container are grouped and deployed at user level.The second is Virtual Container Resource Pools Module,it integrated in the LXC engine used tomanage the cluster hardware resources,To achieve the function of the allocation of resources on demand by generates a simulated container resource allocation environment.The last one is the scheduling module.It’s also integrated in the LXC engine used to monitor the status of different container when it is creating or running.This module aims to explore general laws of the container resources allocation and regulation and implement checkpoint/restore functionality for containers.Finally,Compare and contrast with several other major virtualization technologies,illustrate the performance advantages of the new model in computing power and memory r/w speed.Ultimately prove the rationality and effectiveness of the proposed programs.