Research On Adaptive Deployment Mechanism Supporting QoS In PaXS Cloud Based On LXC

PaaS(Platform as a Service)cloud platform is an integrated and distributed computer system that is composed of the hardware and software infrastructures.The user can use the platform resources configured by PaaS to develop and deploy the application service program,and to manage the execution of applications.LXC(Linux Container)technology is a lightweight virtualization technology that lies on the operating system level,and it brings a new opportunity to construct the PaaS cloud platform.Compared with the service node in the stand-alone environment,the node in the PaaS cloud paltform running environment is quite different.It is an open and very complex operation environment,which is mainly reflected in aspect of platform management activities for its application services,such as analysis,deployment,monitor,and so on.Therefore,if keep deploying application sevices just like the manual operation used in the stand-alone environment,it will be time-consuming and error-prone.Because it needs a long list of complex configuration operations to deploy a service on the cloud platform,and even the experienced developer may cause configuration conflicts at the time of modifying a large number of configuration files,and sequentially result in the service out of normal running.Although many platforms attempt to simplify the configuration process,it still needs the service developer or the PaaS provider to pay out the manual configuration operation.According to the above problems,this paper proposes to build a PaaS cloud platform by using LXC container technology,in order to reduce the platform overheads and enhance its overall performance.What's more,to put forward a self-adaptive deployment mechanism model of supporting QoS based on the front PaaS platform.And this model can select service nodes that satisfy the users' QoS requirement on the basis of the SLA signed between the providers and the users of the cloud platform.Meanwhile,it can deploy the application services rely on the load balance strategy.The specific works are as follows:First of all,on the basis of analyzing the Namespaces and Cgroups mechanisms of LXC,the paper proposes a way to build a simple and lightweight PaaS cloud platform by using LXC virtualization technology for the purpose of isolate the different tenants and share the software and hardware resources of the system.In addition,using the related experiments to prove that the performance of this method is more better compared with the traditional virtual machine method,and it is more suitable for the PaaS cloud platform that to provide the scientific computing service.Secondly,designing the node selection optimizing algorithm of the PaaS cloud platform in order to realize deployment and operation of the applications.Furthermore,creating the objective function on the basis of analyzing the QoS and load balance strategy factors that affect the selection of the platform nodes.It views the QoS attribute values of the current load threshold and the service deployment request of the node as the algorithm's determination conditions.And using the Mixed-Integer Linear Planning to create models as well as to get the solution.This algorithm can realize the deployment task of the application services automatically.Finally,this paper constructs the PaaS cloud platform based on LXC on a server cluster.And to design and realize an adaptive deployment mechanism model that supporting QoS on this platform.And the validity and feasibility of the platform and all the research work have been verified through the system test.Furthermore,to summarize the research contents as well as outlook.