Research And Implementation Of Key Technologies Of Service Composition Based On Cloud Computing

In recent years,with the development of technology and network,people's computing power,resource centralization,resource utilization requirements are getting higher and higher,cloud computing came into being.Different parts of the server's computing and storage capacity is connected together through the network and form a resource pool,and then in accordance with the needs of users,can quickly be allocated to resources or recycling resources.On the one hand,the expansion capability and virtualization features of cloud platform can offer help to effective monitoring and rapid deployment of services in the cloud environment,on the other hand,the type and quantity of the service is increasing,also brings to the cloud platform management inconvenience.In this context,how to effectively deploy,manage and monitor various services,and further combine the services according to the diverse needs of users is the key problem facing the service composition in cloud environment.In the traditional environment,the service selection and composition of research has made some progress.The current study choose the traditional web services as a research object,and a variety of big data services are also emerging,it bring the new challenge to the service composition.Big data services due to its own characteristics,showing different characteristics with the web service.Such as web services' s processing time is relatively stable.But the processing time of big data services increased with the amount of input data.The traditional way of web service calls and composition can not be directly applied to big data services,so the composition of big data services and web services is also a key problem facing the service composition in cloud environment.In order to solve the above problems,this paper designs a service management platform based on Openstack to research and implement the key technologies such as automatic deployment of application,monitoring of application instance,service invocation and service composition based on Artifact.On the basis of this,we also study the QoS-based service composition optimization problem.The method of estimating the running time of big data service using linear regression model and the method of weighting using AHP method is proposed.Finally,we improve the service selection algorithm based on backtracking method,and verify the effectiveness of the linear regression model and improve the performance of the improved backtracking method.