The Design And Implement For Massive Oceanic Geodata Service System Based On Distributed Architecture

With the improvement of information technology, more and more industries, such as telecommunications, water resource, surveying and mapping, finance, weather forecast etc will compute and process the huge data every day. The scale of huge data usually is hundreds of gigabyte (GB) even terabyte (TB). Moreover, the scale of data is growing at index pattern speed rapidly. Such scale of data is called massive data. The effective administration and fast access of massive data is a hot spot issue in present-day.There are two main building schemes for massive data services. Centralized system, it adopts a very high performance server, providing united administration of massive data and united services for users. This scheme's architecture is simple and therefore both the deployment and implement of the system are easily to put into practice. However, the single server always has its limit for the capacity of service performance, and the cost of the system is nonlinear. The expandability of the centralize system is its most weakness. The other scheme is distributed system. It adopts some common servers to administrate the massive data, and the servers share the requests of multi-users. The most advantage of the distributed system is that, it can dynamically adjust the deployment of the servers so as to content the actual requirement. It also can avoid the crash of the system effectively with the benefit of load balance technology. In addition, the cost of the distributed system is less than the centralized system. And the distributed system is adaptive for much more circumstances.The oceanic geodata is massive data. For instance, the scale of the oceanic environment information data of our country is over 300 GB. The service building for massive oceanic geodata is the key point of the digital ocean information base framework.The thesis is proposed a double distributed architecture based on the ArcGIS Server 9.2. Both the map data engines and map service servers are adopted distributed architecture. On the one hand, the map data engines are adopted distributed, it can content the requirement of huge storage but also lighten the pressure of IO. On the other hand, the map service servers are adopted distributed, building the mechanism of the dispatching among the service servers to achieve the load balance, it also used the pooled and the data cached technology.A lot of test results indicated that, the scheme of the system can shorten the response time when multi-users request simultaneously, take full advantage of the system resources, lighten the pressure of the server side, and avoid the crash of the system. In addition, the oceanic geodata of our country is stored the state oceanic administration and its research centers. In the geographic view, the storage of the massive data is distributed and therefore the massive oceanic geodata service building based on distributed architecture is much more feasible.