基于Pi演算的虚拟机动态迁移建模及性能分析

发布时间：2018-03-25 08:32

本文选题：云计算　切入点：虚拟机　出处：《西北大学》2013年硕士论文

【摘要】：随着信息技术的不断发展,云计算作为一种新兴计算模型出现在IT产业。虚拟机迁移技术的不断成熟,促进了云计算的发展。然而,传统的静态资源分配方式存在着资源浪费和资源的非弹性供给等缺点,带有人工参与的资源调整分配方式又有明显的滞后性,以上方法均不能满足云计算环境下对资源的要求,所以本论文致力于研究云计算环境下虚拟机资源的自动调配,实现资源的动态切割和分配,在无需人工干预的条件下有弹性地满足用户的需求。本文结合实验室承担的“云计算及其支撑网络”项目,针对云计算中虚拟机动态迁移问题展开研究,并使用进程代数Pi演算形式化方法对虚拟机迁移前后的服务进行建模及性能分析。首先,在研究了云计算和虚拟机动态迁移相关技术的基础上,设计并实现了一套基于云计算平台的虚拟机动态迁移框架。该框架使得虚拟机能够在无人工干预的情况下,灵活、自主、实时地进行迁移,从而解决了云计算中的故障隔离、能量优化和负载均衡等问题。其次,在完成虚拟机动态迁移实验之后,使用进程代数Pi演算形式化方法对虚拟机迁移前后的服务进行了建模,并利用移动工作台MWB工具对上述建模是否存在死锁、正确性以及系统执行步骤等进行了验证。然后,对虚拟机迁移前后的服务建模从理论上进行了正确性分析。
[Abstract]:With the continuous development of information technology, cloud computing as a new computing model in IT industry. Migration of virtual machine technology continues to mature, to promote the development of cloud computing. However, the static resource allocation in traditional resources wastage and inelastic supply shortcomings, with resource distribution in artificial adjustment there is an obvious lag, the above methods can not meet the requirements for resources in cloud computing environment, so this thesis aims to study the automatic deployment of cloud computing resources of virtual machine environment, realize resource allocation and dynamic cutting, have the flexibility to meet the needs of users in without manual intervention.
Based on the cloud computing and supporting network project undertaken by our lab, this paper studies the dynamic migration of virtual machine in cloud computing, and uses process algebra Pi calculus formalization method to model and analyze the service of virtual machine before and after migration.
First of all, based on the cloud computing and virtual machine migration related technology, designs and implements a cloud computing platform based on virtual machine dynamic migration framework. This framework enables virtual machines to without manual intervention, flexible, independent, real-time transfer, so as to solve the fault isolation cloud in the calculation, energy optimization and load balancing problems.
Secondly, after the completion of the virtual machine migration experiment, service using process algebra Pi calculus of virtual machine migration was modeled, and the use of mobile platform MWB tools for the modeling deadlocks, correctness and system implementation steps were verified. Then, the modeling of virtual machine before and after service the migration of the validity of theoretical analysis.

【学位授予单位】：西北大学
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：TP302

【参考文献】