Research On The Model Of High Available Service Registry In SOC

With the rapid development in Service Oriented Computing (SOC) domain, new problems arise as the booming of existing service quantity. The current service searching mechanisms are able to find services matching the user's query conditions rapidly, but not able to guarantee the availability of recommended services. On the other hand, the service registering with both functionality and quality of service (QoS) are supported by some service registering mechanisms, but the QoS attributes can't be guaranteed during runtime. Both aspects mentioned above will make the service registry less trustable, and have direct negative impact on the availability of the entire SOC environment.In this article, services are modeled using theories of random process and queuing, and more analyze and research is made on the basis of that. A new model of service registry is proposed based on the research job, which is focusing on solving the two main problems in SOC environment– how to guarantee the availability of service searching result, and how to guarantee the runtime service QoS attributes trustable. We can benefit from this model that, not only the rapid and precise searching among massive existing services by using both functionality and QoS attributes as query conditions, but also, the availability and runtime QoS attributes are guaranteed, which is the more significant part. And more, as the result of services'availability guaranteed, the service failure and fall of QoS can be solved by using service replacement within a minimum time period. Since the availability of service registry is guaranteed, the availability of entire SOC environment is improved notably.A stimulation system is also implemented based on the research and design of high available service registry. It is capable of gathering and comparing data of service registering and searching performance under the new model theory, as well as the service runtime performance monitoring and statistic. And also, the process of coping with service failure or fall of QoS is well stimulated in this system.