Technology Study On Cloud For Ocean Data Visualization

In the21st century, the ocean will be the second living place for mankind tosolve the problems of energy, food and water crisis currently which is facing, however,the development of China’s marine economy is still on a weak stage, the marineeconomy accounts only about10%of our country’s total GDP, so it has an importanteconomic and strategic significance for the country to vigorously develop the marineeconomy. To provide enough information and technical support for the developmentof China’s maritime career, National Bureau of Oceanography has implemented a"Digital Ocean" project, and built a large number of marine business applicationsystems. These systems promote the development of China’s marine strongly, butbegin to appear more and more problem in the same time, such as resourceconsumption and high operating costs, lack of system planning and systemmaintenance difficulties. National Bureau of Oceanography intends to adopt theemerging cloud computing and cloud services technology integrating existing marineenvironmental information system to create a efficient marine environment cloudcomputing platform which supplies service on demand for the end-user and savesenergy.Ocean environmental information visualization cloud computing and cloudservices are an important part of the marine environment of cloud computing andcloud services framework application research "(No.201105033) of the NationalBureau of Oceanography’s public projects. Ocean environmental informationvisualization refers to use the theory of high-performance computing and computergraphics and graphic images to show the simulation reproduce or pre now oceanvarious changes in the Intuitive way. Mining the law from the massive ocean data isof great significance for promoting development of marine science and marineeconomic.By analyzing the Demand and task characteristics of ocean data remote interactive visualization, we optimize and transform Hadoop against the problem thatis not suitable for processing remote interactive ocean data visualization byintroducing GPU and MPI into Hadoop to construct a hybrid multi-level granularityparallel computing system firstly, by this way the parallel computing speed of Hadoopis greatly improved; then a mechanism called pipeline is introduced into Hadoop toreduce the data processing delay; through the above measures of optimizing Hadoop,Hadoop can provide low-latency cloud services in the ocean data visualizationapplications. In addition, the mechanisms and the method of realization of theresource layer and application layer is discussed in our three visual cloud computingplatform, such as research and analysis of the vector field and scalar field’svisualization algorithm, XCP platform and so on, eventually a complete ocean datavisualization cloud computing platform is constructed including architecture andimplementation.In order to verify the efficiency and availability of the cloud platform constructedby us, a ocean data visualization cloud platform demonstration system is built, thesystem initially realized the ocean data Remote Interactive Visualization cloudservices. Test results show thatIn order to verify the information of the marine environment, we constructvisual efficiency and availability of the cloud computing platform, we build ademonstration system, the system initially realized the Marine EnvironmentalInformation Remote Interactive Visualization cloud services. By introduce GPU intothe Hadoop parallel computation system. our cloud solution’s cloud service efficiencyis greatly improved. The solution can provide ocean data Remote InteractiveVisualization cloud service, such as browsing, roam, zoom and so on. Above all, thesolution of cloud platform for ocean data Remote Interactive Visualization is a viablesolution.