Research On Key Techonologies Of High-Performance Geospatial Data Service

With the development of geographic information technology and Internet technology,geospatial data services have been used in many domains.There has been more and more attention to the sharing of geospatial data between industries and the sharing of computing power between service systems.For this purpose,the Open Geospatial Information Consortium has established the OGC standard for geospatial data services.Among them,WPS service standard is designed for geospatial data processing.Moreover,there exist some problems in the current geospatial service system.In this thesis,the key technologies for building a high performance geospatial data service system conforming to OGC standard is studied.The main contributions and innovations of this thesis are as follows:First,this thesis gives the overall design of high-performance geospatial data service system.The system has the three-layer structure,namely,Web service layer,application service layer and data storage layer.The Web service layer is responsible for the interaction between application service layer and external system.The application service layer is mainly responsible for the processing of geospatial data.The data storage layer is responsible for storing all kinds of geospatial data and providing access interface for the application service layer.Each layer uses a modular design methodology and is splitted into some modules which can cooperate effectively.Second,the WPS service specification of OGC standard is analyzed,and a general service framework for WPS is built.Three geospatial data processing algorithms,namely slope calculation,aspect calculation and range calculation,are implemented.They are implemented as three specific WPS services.this thesis also defines a simple and flexible extension interface,which facilitates adding new geospatial data processing algorithms to the WPS service without recompiling,making the WPS service highly scalable.Next,to improvce performance effectively,a new load balancing algorithm is designed and implemented.Considering the system fault tolerance,the load balancing algorithm needs to eliminate the presence of the faulty server in the cluster,preventing the server from dispatching the request to servers where there is hardware and software failure.At the same time,taking into account the business characteristics that every request is related to a range of geospatial data,the geospatial data is cached to the servers of the application service layer.Hitting the cache improves the performance of the service,due to preventing from accessing data storage layer.The load balancing algorithm can consider the server resource load and the spatial proximity between request scope and server cache geospatial data.Finally,the test shows that WPS service general framework and specific services achieve the corresponding function and meet the expected requirements,also lays a good foundation for further externsion.At the same time,the test results also show that the load balancing algorithm proposed by this thesis is better than traditional load balancing algorithms such as adaptive algorithm,random algorithm.