Attribute-based filtering for embedded systems

  • Authors:
  • Carlos Mitidieri;Jörg Kaiser

  • Affiliations:
  • University of Ulm, Ulm, Germany;University of Ulm, Ulm, Germany

  • Venue:
  • Proceedings of the 2nd international workshop on Distributed event-based systems
  • Year:
  • 2003

Quantified Score

Hi-index 0.00

Visualization

Abstract

Filters are essential components of a publisher/subscriber communication systems. They provide the necessary selectivity enforcing that a subscriber only is notified about the events for which it actually has subscribed. The paper deals with the problem of establishing a filtering mechanism suitable for distributed systems in which the nodes have memory and performance constraints and the interconnection network has a limited bandwidth, e.g. as in systems composed from smart sensors and actuators. Thus the trade-off between expressiveness on the one side and efficiency and predictability on the other side has to be balanced adequately for coping with the resource constraints. The paper proposes attribute-based filtering which allows to filter on the structural properties of events. The notion of super conformance establishes a relation between these structural properties. Attribute-based filtering constitutes a variant of type-based filtering but reflects a more component-oriented view which is beneficial when dealing with smart sensor or actuators components.