Distributed Data Management And Access Optimization Strategy Of Rendering Application

With the high-speed development of information technology and the increasingpopularity of the Internet, the film and television industry has undergone tremendouschanges. Development from the traditional2D plane film era to the3D digital film eraof sweeping the world, the audiences get the new three-dimensional experience from thevisual effects. And one of the key technologies of achieving that effect is the rendertechnology. However, with the increasing expansion of the information currently thetraditional stand-alone rendering application under has been unable to meet the demand.With the rapid development of distributed computing distributed rendering applicationsare increasingly spread widely. The strong computing power of cluster brings a hugeupgrade of the performance of rendering. But at the same time frequent data accessmakes the disk IO become the bottlenecks of performance improvement of distributedrendering application.The system of distributed data caching makes rendering application to store all thedata which is created in the process by the way of building a distributed share memoryregion. In the efficient distributed memory resource management module the memorypool status information management mechanism is responsible for real-time monitoringthe information of the shared memory. Making use of memory data managementmechanism to record metadata of all the information stored in the shared memory andfacilitate quickly to find their locations. The rendering uses can freely use memoryresources from the memory resource pool. Meanwhile in the conventional distributeddata access method, based on the characteristics of the system for the rendering thearticle presents two data access optimization strategy. Double multi-granularity lockingmechanism is proposed to not only avoid conflicts of multiple rendering user access todata but also effectively reduce the overhead of waiting. Data prefetching optimizationstrategy is presented as the user’s pretreatment process before reading data. It effectivelyreduces network overhead by the way of avoiding the repetition data request of the samestorage node and improves the efficiency of data access. Experimental results show that for distributed rendering application distributed datacaching system can well meet the need of data storage and access of the renderingapplication. Complete distributed memory data management provides reliable dataprotection for the rendering application. Optimization strategies of data access not onlymeet the requirements based on the rendering applications but also bring theimprovement of performance of data access.