Sensor based transaction temporal database architecture

Abstract

Sensor data processing and managing is a fundamental requirement of industrial systems. Validity of these data is time bordered and should be stored in database. However, existing database solutions are inadequate in terms of performance - effectiveness of the whole system, which is manifested using the size of the managed data and processing time. This paper deals with the principles of extended temporal data modelling based on the attribute level, which is very useful for time-various sensor data processing. It deals with the transaction management problems and describes solutions. However, each sensor has defined measurement precision, thus only reliable data based on significant change are stored. This concept is mostly designed for industrial communication systems, sensor data processing, intelligent transport systems, but also for medical applications.