Under The Distributed Parallel Computing Gml Spatial Data Replication Synchronization Update Mechanism Research

As a geographic information sharing standard, GML’s storage and management theory in single server have become increasingly mature at present. There have been related domestic research achievements, but in distributed storage parallel computing environment theory, it still has no siginificant results.The distributed parallel computing concept has exerted at home and abroad, before a few years in academic circles rise of Grid Computing, today’s popular Cloud Computing, Global related network and Intelligent City and so on are involve to distributed storage and parallel computing areas. As the knowledge discovery and technical progress, more and more data need spatial to make3D schema, even adding time dimentionality to beome4Ds to present before common eyes. These data is quite large and single server can not adapt to the large amount of data storage treatment requirements, but also related to the spatial data format, so it is inevitable to research distributed storage huge data and high performance parallel Parallel computing processing GML.Based on the related storage management environment GML in single server, The article research GML spatial data synchronization replication and update in distributed parallel surroundings. Then make distributed opensource database-MongoDB as experiment and test, This mainly contains4points as follows:(1) Distributed GML storage, puts forward the concept of distributed GML spatial database and3design schemas. Opensource NoSQL database MongoDB plays a roll in GML distributed storage and Index, Query and so on, then puts forward the theory of Cloud GML management and verification is gived in building distributed condition.(2) Parallel GML synchronization replication strategy, explain traditional spatial database and distribute spatial database, elaborage the architecture and theory of distributed database replicas, then make ArcGIS10geodatabase version synchro replica as reference, propose distribute GML "Feature Grouping" synchro replica strategy. It can deal with the problem of huge GML datum when dividing data in distributed cluster,and the conflict and solutions are put forward at last.(3) Distributed GML datum update mechanism, data updating should bring about parallel access and consistency preblem. Starting from distributed transaction management, describe the key technology of consistency update, analyse ArcGIS10long transaction and topology mechanism. Then discuss GML geometry and topology consistency update, put forward distribute "GML document update model", enrich distributed GML spatial data update theory.(4) MongoDB GML replication and update design, According to the proposed GML syschronication replica and update strategy,build the distributed parallel computing environment with MongoDB,then test the efficiency of GML datum syschronication replica and update.And the distributed tesed tool-JMeter is used to test the circumstance;Finally, through the distributed parallel computing environment and research to spatial data storage management of GML at the bottom,this article makes a deep research on its distributed data replication and update efficiency mechanism. So it can develope a slice of heaven and earth for GML in distribute and parallel computing field.