W.G. 4

Title: Resources Other Than Time

Working Group 4 Leader: 

Prof. Stefan M. Petters

Instituto Superior de Engenharia do Porto
Rua Dr. António Bernardino de Almeida 431
4200-072 Porto, Portugal
Telephone: +351 228340502 (Ext: 0415160)
e-mail: smp <at> isep.ipp.pt
 

Description: For embedded systems design, it is paramount that timing analysis be reconciled with other resource constraints. This is best done within models that permit the exploration of trade-offs between multiple dimensions, such as functionality, performance, quality-of-service and resource consumption. To support compositionality, this ability must be carefully balanced against the need to separate concerns as much as possible. This Working Group aims to find formalisms that are appropriate for different purposes, such as time-energy trade-offs in energy-constrained computing. In order to support component-based design, the relevant dimensions must then, ideally, be captured within interfaces. The timing analysis community has developed sophisticated analysis methods which likely can be adapted to also analyse other resources. This Working Group will interact with the other concerned research communities to achieve this. This Working Group will address:

  • Modifying timing analysis techniques to work for resources such as energy dissipation and memory. For energy, this could be achieved by further development of the priced timeautomata model and combining techniques from model-checking and static analysis, or by using amortised analysis. Secondly, probabilistic methods will be deployed to gauge the expected spare capacity in the system. The results of both approaches will be exploited into e.g., optimal energy-aware scheduling strategies.
  • How analyses for different kinds of resources can support multi-objective optimisation. One line of research for this will involve the continuation of existing work on multi-weighted automata and on multi-weighted modal transition systems with structured labels - allowing for analysis of multi-objective optimisations.
  • Multi-objective optimisation during code generation through integration of more detailed energy models for actual processor architectures to WCET-aware compilers and through twophase optimisations.