Resource-Aware Scheduling of Distributed Ontological Reasoning Tasks in Wireless Sensor Networks

Abstract

Abstract-As the number of wireless sensor network applications continues to grow, the need for specialized task scheduling mechanisms, aware of the sensor devices' capabilities and realtime resource availability, is becoming more and more apparent. In this paper, we therefore propose a generic model for task scheduling in heterogeneous networks, which we subsequently use to schedule distributed reasoning tasks, originating from a real-world WSN monitoring and management application. By means of simulation, we evaluate several developed scheduling heuristics and compare the results to an optimal solution of the same WSN task scheduling problem, obtained using ILP. Experiments show that our heuristics produce acceptable task schedules while maintaining a low resource footprint.

On the design of a flexible software platform for in-building OTT service provisioning

“…[15] Our own bin-packing-based algorithms have been shown to produce acceptable solutions. [16] III. SOFTWARE ARCHITECTURE Figure 1 shows a high-level architectural view of the platform we propose.…”

“…• Scheduling the deployment of OTT services is done using algorithms based on the bin packing metaphor; these were previously published in [16].…”

“…The measurements above formed the basis of a test case for the scheduling heuristics we published in [16]. We used time slots representing 10 seconds of wall clock time.…”

“…Like in [16], the Best Fit heuristic lives up to its name and slightly outperforms its peers. For our type of problem, Fair Fit appears to be the worst choice, both in terms of combined execution times and node occupancy; we do however note that its time loss of 8.73% is still a fairly competitive result.…”

Autonomous Service Composition in Symbiotic Networks

“…Building upon this model, we applied bin packing heuristics to the problem of scheduling tasks in heterogeneous environments, thereby providing a preliminary answer to research question 3. Both the model and the scheduling algorithms are detailed in [4].…”