Designing And Implementing A Distributed Operation System Based On Akka

As the massive data explodes,increasing distributed platforms are being used in large-scale clusters.Meanwhile,with the platform is common used in commercial applications,the instability of distributed platform in the complex cluster environment has gradually highlighted.For economic and benefit reasons,enterprise server clusters usually cannot guarantee that all nodes have the same software and hardware configuration in the real industry environment.The diversity of this configuration will not only bring idle waste of node resources,but also interference during the normal operation of the server cluster and resource scheduling.At the same time,internal problems of the cluster will also interference the normal operation of the server.All these disturbances have become obstacles to the large-scale commercialization of distributed platforms.How to quick diagnose these interference sources becomes a problem that needs to be solved urgently.In order to diagnose the fault source which affects the operation of the distributed platform,this thesis designs a distributed operation system based on Akka.The distributed design system can achieve rapid detection of cluster failure and help users solve the problem.First of all,this thesis starts from the user needs.the effectiveness of the system is guaranteed by user requirements analysis.Secondly,the general function structure and framework of distributed operation system are introduced.The distributed operation system is divided into three modules:detection module,user interface module and management module.Among them,the detection module is responsible for the testing work,including network environment testing,platform environment testing and system environment testing.The user interface module provides the way the system interacts with the user,including the web page interface and the script interface.The management module is the control center of the whole system,manages the node information in the cluster,the command protocol and the task operation.The management module uses high-reliability Akka technology to build the M/S architecture and manage all the node.Finally,the distributed operation system is implemented.The prototype system performs the corresponding detection or query function according to the user's input,and displays the result to the user.The experimental results show that the distributed operation system is effective and stability performance.