Skip to Main content Skip to Navigation

Declarative approach for long-term sensor data storage

Manel Charfi 1, 2 
2 BD - Base de Données
LIRIS - Laboratoire d'InfoRmatique en Image et Systèmes d'information
Abstract : Nowadays, sensors are cheap, easy to deploy and immediate to integrate into applications. These thousands of sensors are increasingly invasive and are constantly generating enormous amounts of data that must be stored and managed for the proper functioning of the applications depending on them. Sensor data, in addition of being of major interest in real-time applications, e.g. building control, health supervision..., are also important for long-term reporting applications, e.g. reporting, statistics, research data... Whenever a sensor produces data, two dimensions are of particular interest: the temporal dimension to stamp the produced value at a particular time and the spatial dimension to identify the location of the sensor. Both dimensions have different granularities that can be organized into hierarchies specific to the concerned context application. In this PhD thesis, we focus on applications that require long-term storage of sensor data issued from sensor data streams. Since huge amount of sensor data can be generated, our main goal is to select only relevant data to be saved for further usage, in particular long-term query facilities. More precisely, our aim is to develop an approach that controls the storage of sensor data by keeping only the data considered as relevant according to the spatial and temporal granularities representative of the application requirements. In such cases, approximating data in order to reduce the quantity of stored values enhances the efficiency of those queries. Our key idea is to borrow the declarative approach developed in the seventies for database design from constraints and to extend functional dependencies with spatial and temporal components in order to revisit the classical database schema normalization process. Given sensor data streams, we consider both spatio-temporal granularity hierarchies and Spatio-Temporal Functional Dependencies (STFDs) as first class-citizens for designing sensor databases on top of any RDBMS. We propose a specific axiomatisation of STFDs and the associated attribute closure algorithm, leading to a new normalization algorithm. We have implemented a prototype of this architecture to deal with both database design and data loading. We conducted experiments with synthetic and real-life data streams from intelligent buildings.
Document type :
Complete list of metadata
Contributor : ABES STAR :  Contact
Submitted on : Friday, November 30, 2018 - 3:49:10 PM
Last modification on : Tuesday, June 1, 2021 - 2:08:08 PM
Long-term archiving on: : Friday, March 1, 2019 - 2:57:50 PM


Version validated by the jury (STAR)


  • HAL Id : tel-01940920, version 1


Manel Charfi. Declarative approach for long-term sensor data storage. Databases [cs.DB]. Université de Lyon, 2017. English. ⟨NNT : 2017LYSEI081⟩. ⟨tel-01940920⟩



Record views


Files downloads