延迟光照算法对Hadoop集群渲染系统的改进与实现

Improvement and implementation of Hadoop cluster rendering system based on deferred lighting algorithm

针对集群渲染系统在三维场景文件输出过程中,海量数据处理以及存储的问题,利用Hadoop框架对现有系统进行分布式模型改造,以基于延迟光照算法的处理方式,对分布式存储和计算的业务逻辑结构部署,实现改进Hadoop集群渲染系统对任务的处理策略,提供高容错、可靠的存储和海量计算数据的方式。通过实验证明改进的Hadoop集群渲染系统可以提高系统利用率和任务处理能力。

In view of the problem of massive data processing and storage in 3D scene file output process in cluster rendering system, the distributed model is reconstructed for the existing system by using the Hadoop framework. It uses deferred light- ing algorithm to realize the business logic structure of the deployment of distributed storage and computing. The task process- ing strategies to improve Hadoop cluster rendering system is provided with high fault tolerance, reliable storage and mass data calculation. Experimental results show that the developed Hadoop cluster rendering system can improve system efficiency and tasking ability.