The Architecture Design And Implementation Of Software Self-adaption Deployment Based On Cloud

In recent years,the computing services are inseparable from the support of cloud computing with the rapid development of cloud computing.Cloud computing is an information resource service system based on Internet.Users can get the super computing resources for their own services via the Internet.The core idea of cloud computing is the integration of physical computing resource and the use of virtualization technology to achieve resource sharing.Cloud data center builds a virtual resource pool with computing,storage and network resource to achieve the deployment automation and dynamic allocation of cloud resources.How to manage such a large scale cloud resource scientifically has become a new challenge along with the increasing data.The idle resource and low resource utilization of cloud data center results in a large waste of cloud resource.To solve this problem,this thesis focuses on the dynamic scheduling of cloud resource,designs and implements a software self-adaption deployment architecture based on cloud.Hadoop is an open-source Java-based distributed cloud platform for data analysis and processing,which promotes the development of Big Data by the dual advantages with high capacity and low cost.This theis select Hadoop as the experiment basis.The applications are developed with some specific purpose to provide convenience for lives.And the evolution of cloud computing provides guarantee to the applications.Deploying the applications on the cloud makes the high concurrency and high performance in applications become possible.Therefore,it is a valuable research to find out how to use cloud resource wisely and to implement cloud resource dynamic scheduling when the application load changed.This thesis selects existing applications in our laboratory to study the software self-adaption deployment based on cloud.The main contents of this thesis are how to build and use Hadoop,how to implement the real-time monitoring in running status of cluster nodes and the monitoring data reading by program,how to design rational demand forecasting of resources with changing load according to application load and the usage of cluster resource,and studying the deployment and migration technology of cloud virtual machine to schedule the virtual resource.Finally,desiging and implementing the software self-adaption deployment architecture based on cloud.The self-adaption deployment architecture is designed to improve applications' service capabilities by cloud computing ability,Qos standard and load balance.Meanwhile,when the application load changed,the architecture can percept the change and make demand forecasting with it.And the resource scheduling requests,that is releasing resource while low load and creating resource while high load,can be send according to the forecasting.In this way,the cloud idle resources and waste issues are effectively solved,and the resource utilization is elevated.