Links to document based files:

• Entire Document (89 pages, 2564 KBytes)
• Document in pieces
  • Abstract, Acknowledgements, Table of Contents, and List of Figures (9 pages, 9 KBytes)
  • Chapter 1: Introduction (10 pages, 78 KBytes)
  • Chapter 2: Review of Related CASE Tools (10 pages, 65 KBytes)
  • Chapter 3: Real-Time Scheduling Theory Background (16 pages, 41 KBytes)
  • Chapter 4: Theoretical Foundations of AFTER (20 pages, 38 KBytes)
  • Chapter 5: Implementation of AFTER (16 pages, 2351 KBytes)
  • Chapter 6: Summary and Future Work (4 pages, 5 KBytes)
  • References (4 pages, 6 KBytes)

Abstract

During the post-implementation stage of the software lifecycle of an embedded system, the developers are faced with the task of fine-tuning the system timing parameters. They are forced to resort to time consuming trial-and-error debugging methods. Our goal is to create an automated tool that facilitates the post-implementation debugging process. The main focus of this tool, which we call AFTER (Assist in Fine-Tuning of Embedded Real-time systems), is to detect anomalies with the system timing parameters and assist thedeveloper in finding a solution for these problems in a systematic manner. AFTER uses raw timing data collected from the embedded system, analyzes it, then provides a temporal image of the current implementation, highlighting actual and potential problems. It can also be used in the predictor mode, where it allows the user to interact with it and assist him in narrowing down the exact fine-tuning optimizations required to solve the timing problems of the target system.

The key contributions of this work are: combining the real-time scheduling theory of various research groups and extending it to create mathematical models for analyzing embedded real-time systems; creating an automated analysis and predictor tool that assists in fine-tuning of embedded real-time systems after implementation in a systematic manner.