Method and device for creating earthquake big data storage application model

Abstract

The invention relates to the field of petroleum seismic exploration, in particular to a method and a device for creating an earthquake big data storage application model, seismic data is cut into three parts, namely a file header, a data channel header and a data body, the data body is cut into data blocks according to requirements, and general file information is stored in a database; the data body is stored in a disk array, and data channel header information and index information are stored in a file system; the trace header keyword and the data body of the data information are separately stored, so that the flexibility of data calculation can be better reflected, the reading and writing times of the data are reduced, the index construction efficiency is greatly improved, the universality and the cross-platform type of seismic data format application are realized, a rapid and efficient index construction model is supported, and the dimension reduction storage of mass data is realized; distributed data reading and writing can be supported, and the feedback speed of the disk array in the data reading and writing process is greatly increased; by applying the local I / O cache, the data read-write efficiency is improved.

Description

technical field [0001] The invention relates to the field of seismic exploration, in particular to a method and device for creating a seismic big data storage application model. Background technique [0002] With the continuous progress of petroleum exploration and acquisition technology, the acquisition density and accuracy of exploration data continue to increase, and the volume of data continues to increase, and the volume of collected data reaches PB level. The ever-increasing amount of data has caused many bottlenecks in the storage, retrieval and application of exploration data. The main issues are: [0003] Difficulty in storage: PB-level seismic data volumes are loaded into cluster parallel machines at the same time. Conventional sequential data parsing and reading and writing are relatively slow, which greatly affects the cycle and efficiency of data processing implementation. [0004] Difficulty in retrieval: Retrieving the storage location of the target data fro...

AI Technical Summary

Problems solved by technology

There are problems in the conventional retrieval model based on the database: First, the PB-level data volume index in the database occupies a lot of storage resources, and the storage disk capacity requirements are very high; Class retrieval efficiency is unstable

[0005] Difficulty in reading and writing: In the process of reading and writing, the data requirements are divided into two types. One is to calculate by channel, and the data requirement for each time is KB; , the conventional mode of reading and writing data according to data requirements will cause too many read and write requests sent to the disk array to cause read and write congestion, which will affect the I / O efficiency of data, and small-scale data read and write cannot be fully applied Existing 10 Gigabit bandwidth

[0006] For the above problems, no effective solutions have been proposed so far

Images