基于流量检测的目标大数据快速检索系统设计

Design of fast retrieval system for target big data based on traffic detection

目前使用的基于簇内乘积量化、倒排索引的检索方法存在耗时较长、效率偏低的问题。为此,设计了基于流量检测的目标大数据快速检索系统。设计索引服务器,利用XML驱动系统,编写XML配置文件,并传递给接口执行相关操作。使用流量检测方法检测异常数据入侵行为,实现对异常数据的定位。计算采样间隔,并确定索引构建时间,构建基于时间域的倒排索引。设计检索流程,完成系统设计。实验结果表明,该系统索引构建时间最短为0.03 s,膨胀系数最大可达到0.99,具有快速检索目标大数据的效果。

The current retrieval methods based on intra cluster product quantization and inverted index are time⁃consuming and inefficient.Therefore,a fast target big data retrieval system based on traffic detection is designed.Design the index server,use the XML driven system,write the XML configuration file,and pass it to the interface to perform relevant operations.The traffic detection method is used to detect the intrusion of abnormal data and locate the abnormal data.Calculate the sampling interval and determine the index build time.Build an inverted index based on time domain.Design the retrieval process and complete the system design.The experimental results show that the minimum index construction time of the system is 0.03 s,and the maximum expansion coefficient can reach 0.99,which has the effect of fast retrieval of target big data.