基于分级存储的石油勘探地震处理应用研究

石油勘探中“两宽一高”采集技术的深入应用,对地震资料并行处理能力提出了更高的要求,针对存储对地震资料处理应用的影响进行研究。为提高各类应用的运算效率,设计并实现了基于SSD、集群NAS多级分层存储地震资料处理方法,针对高IOPS应用、大数据块并行应用进行算法流程级设计。通过DD与IOZONE基准测试、两种应用实际运行效果,验证了设计方案具备在地震资料处理中应用效率与性价比上的优势,适用于勘探数据中心地震资料处理,具有一定的普适性。

Research on Seismic Data Processing Based on Multi-Level Storage

As the technology of “Two Width and One Height” in petroleum exploration applied in depth, it makes a higher demand for seismic data parallel processing capacity, this paper focuses on the impact of storage on seismic data processing applications. To improve the efficiency of various types of applications, it designs and implements multi-level hierarchical storage for seismic data processing method based on SSD and cluster NAS, which designs algorithm flow for high IOPS applications and parallel applications contain large data block. It proves that the design has the advantages of application efficiency and cost performance in seismic data processing, which uses the DD and IOZONE benchmark test and two applications running result. The design is suitable for the seismic data processing of the exploration data center, and has certain universality.