Cross-system Optimization of Data-stream Queries

Cross-system Optimization of Data-stream Queries

(Third Party Funds Group – Sub project)

Overall project: FOR 1508: Dynamisch adaptierbare Anwendungen zur Fledermausortung mittels eingebetteter kommunizierender Sensorsysteme
Project leader: Klaus Meyer-Wegener
Project members: Sebastian Herbst
Start date: 01/08/2012
End date: 31/12/2018
Acronym: BATS-TP3
Funding source: DFG / Forschungsgruppe (FOR)
URL: http://www.for-bats.de/

Abstract

In resource-limited sensor networks an efficient data collection is particularly important.  For that, a distributed data-stream system is used, which has a query language that con activate and control the following four components:  The first is a central data management, which is created together with the sub-project 4 on network management and which can hide the heterogeneity of the sensor network.  The mobile and ground-based sensor nodes can have varying software and energy states.  Because of the resource limitations, the applications are distributed across several nodes.  The second component is an abstract model of the evaluation (a query), which is easy to create and still leaves lots of room for the implementation.  The query is model-based and thus can be transformed in many ways and can be executed efficiently.  The third component is a cost model for a quantitative assessment of datastream queries, e.g. with respect to energy efficiency.  Different optimization goals can be imagined here, for instance referring to the accuracy of flight trajectories.  And the fourth component finally consists of newly developed special operators and a description of the required accuracy, which allow for an advanced optimization to extend battery lifetime.

Publications

• Herbst S., Tenschert J., Meyer-Wegener K.:
  Using Data-Stream and Complex-Event Processing to Identify Activities of Bats
  BTW 2015 (Hamburg, 02/03/2015 - 03/03/2015)
  In: Datenbanksysteme für Business, Technologie und Web - Workshopband, Bonn: 2015
  URL: https://www6.cs.fau.de/publications/public/2015/btw15_herbst.pdf
  BibTeX: Download
• Herbst S., Pollner TN., Tenschert J., Lauterwald F., Endler G., Meyer-Wegener K.:
  An Algebra for Pattern Matching, Time-aware Aggregates and Partitions on Relational Data Streams
  DEBS Distributed Event-based Systems (Oslo, 29/06/2015 - 03/07/2015)
  In: DEBS '15 Proceedings of the 9th ACM International Conference on Distributed Event-Based Systems, New York, NY, USA: 2015
  DOI: 10.1145/2675743.2771830
  BibTeX: Download