Software Architecture for a Multi-Level Real-Time System

Abstract

When a real time system has a mixed set of time critical tasks, including tasks with hard deadlines and tasks with soft deadlines, managing a mixed set of tasks in a timely manner becomes harder and requires a multi-level architecture. This thesis concentrates on building such an architecture. The proposed architecture is based on the current Computer Aided Prototype System (CAPS) architecture, which only deals with hard real time and non-time-critical tasks. Priority based scheduling techniques are used along with Ada tasking to schedule different levels of tasks. Periodic hard real time polling tasks are used to insert sporadic soft real time tasks into the system. A method is developed to assign deadlines to soft real time tasks dynamically. Two tasking packages are added to the system for scheduling and execution of soft real time tasks. The Earliest Deadline First (EDF) algorithm is used dynamically to schedule soft real time tasks. A pilot prototype is developed to test the proposed architecture via a run time monitoring package. The results show that the proposed system guarantees that all hard real time tasks meet their deadlines and an acceptably small percentage of soft real time tasks miss their deadlines.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Sep 01, 1998
Accession Number
ADA357212

Entities

People

  • Omer Korkut

Organizations

  • Naval Postgraduate School

Tags

Communities of Interest

  • C4I
  • Energy and Power Technologies
  • Ground and Sea Platforms
  • Materials and Manufacturing Processes

DTIC Thesaurus Topics

  • Algorithms
  • Computer Programming
  • Computer Science
  • Computers
  • Control Systems
  • Debugging
  • Engineering
  • Guarantees
  • High Level Languages
  • Models
  • Operating Systems
  • Programming Languages
  • Prototypes
  • Scheduling (Production)
  • Software Design
  • Software Development
  • Software Prototyping

Fields of Study

  • Computer science
  • Engineering

Readers

  • Parallel and Distributed Computing.