With the development of new exploration technology,the application of machine learning is increasing,the volume of seismic data is increasing,and the storage resources are increasing.As the basic technology of database,cold and hot data separation technology divides data into hot data and cold data.Based on the cache replacement strategy,combined with LRU algorithm and LFU algorithm,and according to the seismic data processing law,a cold and hot separation strategy for seismic data processing in machine learning is designed.The application of cold and hot data separation technology on seismic data can effectively reduce the cost of storage hardware,reduce the storage burden of HPC cluster,and improve the efficiency of data processing.
关键词
冷热数据分离/地震数据处理/缓存替换策略
Key words
cold and hot data separation/seismic data processing/cache replacement strategy
维普
万方数据