Method for storing and searching mass sensor data

Abstract

The invention discloses a method for storing and searching mass sensor data and belongs to the field of sensor data. The method comprises the following steps that: two layers of storage frame, which comprise a root node and a plurality of leaf nodes, are established; a receiving server distributes sampling data sent by a sensor to the corresponding leaf nodes; each leaf node stores the historic sampling value of the same monitoring target, which is acquired by the sensor, in an element group of a leaf node database in a uniform form to form a sampling data sequence of the monitoring target; and an overall time and space distributed index and an overall key word distributed index of the two layers of storage frame are established. Quick time and space search and key word search can be supported at the same time through the overall key word index, the overall time and space index and an overall search processing mechanism. Multiple searching types of quick processing are compatible, and the defects that the current time and space database is mainly used for storing relative static space data or relative single time and space data and an efficient heterogeneous data flow representation method is lacked are overcome.

Description

technical field [0001] The invention relates to the field of sensor data and the field of temporal and spatial databases, and proposes a massive sensor sampling data storage and query processing method capable of simultaneously supporting fast space-time query and keyword query. Background technique [0002] At present, in various sensor application systems, the demand for centralized storage management and query processing of massive sensor sampling data is very strong. For example, in Internet of Things systems, intelligent transportation systems, and urban real-time monitoring systems, it is necessary The continuously dynamically uploaded sensor sampling values ​​are centrally stored in the data center, and the history and current sampling values ​​of the sensor are quickly queried and analyzed through the corresponding query interface, so as to carry out corresponding processing. [0003] However, there is currently no effective solution to the above requirements. In te...

AI Technical Summary

Problems solved by technology

However, this method also has many defects when dealing with sensor sampling data management.

First of all, the performance of relational database clusters in processing keyword queries is much lower than that of key-value databases, and cannot well support "key-value" queries; secondly, relational database clusters adopt strict transaction processing mechanisms (such as data locking mechanisms , multi-version mechanism, etc.), so in the face of frequently updated sensor sampling data, the efficiency is very low, which greatly limits the scale of sensor data that can be supported; finally, relational database clusters usually only support conventional general data types and Query operations, but cannot effectively represent, calculate and query processing complex heterogeneous sensor sampling data with spatio-temporal attributes

Images