Loading...
Search
Search in this resource
sort by
Send to Friend
 Digital Object List
 Bookmark
 Review List
  1. Simple search
  2. Search result
  3. کاهش سربار انرژی مصرفی ناشی از روش بازگشت از خطای عقب گرد در سیستم های نهفته توزیع شده

Reducing the Energy Overhead of Rollback/Recovery Mechanisms in Distributed Embedded Systems

Khavari Tavanaو Mohammad | 2010

899 Viewed
  1. Type of Document: M.Sc. Thesis
  2. Language: Farsi
  3. Document No: 41209 (19)
  4. University: Sharif University of Technology
  5. Department: Computer Engineering
  6. Advisor(s): Ejlali, Alireza
  7. Abstract:
  8. Hard real-time embedded systems often operate in harsh environmental conditions that necessitate fault tolerant computing techniques. These embedded systems are also energy-constrained, since energy consumption is one of the key objectives. Examples include satellites, surveillance systems, autonomous seaborne and airborne systems. Generally, fault tolerance is obtained by hardware or time redundancy in these systems. Checkpointing with rollback recovery which employs time redundancy is one of the common techniques for tolerating transient faults in real-time embedded systems. Using checkpointing in energy constrained hard real-time systems should be carefully employed. In this thesis, we propose a novel checkpoint placement policy for low-energy and k-fault tolerant embedded systems which dramatically decreases the number of checkpoints. This scheme which is called quasi-static (QS) checkpointing composed of two parts: Static part, and dynamic part. Static part performs at design time and finds feasible and non-uniform checkpoint intervals, and when there is no fault, the system uses these intervals. Once a fault has been detected, the system uses dynamic part to determine checkpoint intervals based on optimal equidistant checkpointing (OPT). Simulations show that the proposed scheme can decrease execution time up to 19.6% compared to optimal equidistant checkpoint policy. In addition, this scheme does better than OPT in finding feasible checkpoint placement for tasks with high utilization. Finally, we have extended a system energy model for checkpointing system and make a comparison among some other approaches from energy consumption point of view. Also, we propose a novel feedback based approach for standby sparing scheme which guarantee one fault per task in the system. In this system primary processor exploit DVS and standby processor use DPM technique to decrease energy consumption. Based on the feeadback of actual execution time, voltage level of the tasks are chosen in the such that meet timing constraints of the hard real-time systems. We compared suggested scheme with other low energy standby sparing and show the effectiveness of the feedback based system
  9. Keywords:
  10. Embedded Real-Time System ; Checkpoint ; Proportional-Integral-Derivative (PID)Controller ; Hard Real Time Systems ; Rollback Recovery ; Standby Sparing ; Feedback System

 Digital Object List

 Bookmark

No TOC