Design And Implementation Of A Service Governance Platform Based On Container Cloud

In recent years,cloud computing technology has undergone great changes.In business community and open source community,there are many cloud platforms built on container technology,such cloud platforms are called container cloud.Micro-service software architecture breaks down a software system into separate subsystems,those subsystems are called services,each service generates instances on demand and always deployed in container.The purpose of service governance is to ensure that the capabilities of the micro-service software architecture are fully realized,which is an indispensable feature for the container cloud.To complete the service governance task in container cloud,a service governance platform has been designed and implements in container cloud.The main objects of service governance in container cloud are those software systems which deployed on container cloud,using micro-service software architecture.Considering the characteristics of this object,this platform consists of four modules: service description,service instance registration and discovery,service monitoring,and service load balance.They realize four functions such as service description,service registry and discovery,service running status data collection and service load balance.To complete this service governance platform,the main works are listed: Expanded existed service description method for the service which using messaging queue based on the existing web service description standard,Extended the scope of the use of the platform.Referred to the existing service governance platform and considering the features of container cloud,redesigned a service registration and discovery center,compared to other implementations,the record of the relationship between the service instance and the Docker container,Docker container and host is added.Monitored each service instance by monitoring each Docker container,including CPU rate,memory usage,disk and network input/output usage and rate.In addition to the support of classic load balance algorithm,also implemented a dynamic load balance algorithm,the algorithm adjusts the weight of each node by interacting with their monitoring data,making the allocation of all load more reasonable.By comprehensive test,finally proved that the service governance platform can correctly describe the micro service system,get the status of service instances and to ensure the micro service system which being governed more reliability and efficiency.