Loading...

On-Line Energy Management for Fault-Tolerant Real-Time Embedded Systems

Aflaki Khorasgani, Mehdi | 2012

598 Viewed
  1. Type of Document: M.Sc. Thesis
  2. Language: Farsi
  3. Document No: 44271 (19)
  4. University: Sharif University of Technology
  5. Department: Computer Engineering
  6. Advisor(s): Ejlali, Alireza
  7. Abstract:
  8. Providing low-power fault tolerance is usually an important requirement in real-time embedded systems. A faulty result in safety-critical systems and a failure to meet a deadline in hard real-time systems can cause catastrophic consequences for the system or environment. Therefore, these systems must be highly reliable and meet real-time constraints. Slack time in time redundancy techniques (e.g. roll-back recovery) can be used to increase system reliability and to save energy. Time redundancy techniques are suitable from an energy consumption point of view but may not be sufficiently reliable for tight real-time constraints. In the condition of tight timing constraints, the hardware redundancy techniques can be used to provide reliability requirements. However, adding redundant hardware components can impose significant energy consumption to the system. Energy management methods can be used to restrain of excessive energy consumption of hardware redundancy techniques. There is an interesting trade-off between system reliability and energy consumption. In this thesis, we used a combination of standby-sparing and time redundancy techniques to provide both high reliability and low energy consumption. For the system we also proposed an online energy management method to exploit available dynamic slack. We compared our proposed technique with one of the most recent online energy management methods on standby-sparing system. The results show that the proposed scheme provides much better energy saving while maintains the original system reliability
  9. Keywords:
  10. Energy Management ; Embedded System ; Real Time System ; Fault Tolerance ; Embedded Real-Time System ; Static Slack Time ; Dynamic Slack Time

 Digital Object List

 Bookmark

No TOC