Object Structures for Real-Time Systems and Simulators

1
Object Structures for Real-Time Systems and
Simulators

• K.H. (Kane) Kim
• Presented by Denh Sy
• 8/3/06

2
K.H. (Kane) Kim

• Professor of Electrical Engineering, Computer
  Science and Informatics at University of
  California Irvine
• Directs the UCI DREAM (Distributed Real-time Ever
  Available Microcomputing) Lab
• IEEE Computer Society's 2004 Tsutomu Kanai Award
  winner

3
Problem

• Building and simulating modern complex real-time
  distributed applications
• Current tools programming languages and
  architectures are still deficient
• Usually to achieve good response time avoid
  software program hardware

4
Problem

• Building interactive systems of heterogeneous
  objects leads to weak traceability
• Requirements (temporal or behavior) are not
  rigorous
• Design techniques are not integrated timing,
  execution capabilities, fault tolerance and
  security

5
Solution

• Time concerns and more importantly interactivity
  of different objects makes conventional
  construction of systems and simulators
  problematic
• TMO (Time-triggered Message-triggered Object) is
  a design philosophy used to construct such
  systems allowing for increased productivity,
  ignoring the minutia of construction

6
TMO Object structure

• Object data store name, capabilities, etc.
• Service methods a way of handling messages and
  requests
• Real-Time clock needed

7
TMO

• Two new features
• Spontaneous methods time triggered methods
  invoked when real-time clock reaches a specified
  time
• Basic concurrency constraint Giving time
  guarantees on execution times and proper
  execution of Spontaneous Methods (SpM) and
  Service Methods (SvM)

8
Why is this important

• Real-time programming in embedded devices
  becoming essential
• Need fast development of software
• Software should be reliable and perform well

9
Using TMO

• He designed an CAMIN (Coordinated Anti-Missile
  Interceptor Network)
• Which distinguishes hostiles and non-hostiles
• Using only high-level specifications he was able
  to set up the design for CAMIN

10
(No Transcript)
11
Stated advantages

• Strong traceability between requirement
  specification and design
• Cost-effective high coverage validation
• Autonomous subsystems and ease of maintenance
• Flexible framework for requirement specification

12

• Seems to introduce some aspects of fuzzy logic in
  SpM timing
• TMO package is still in nascent stage, though a
  comprehensive package is due for delivery soon
• Design philosophy is thoroughly presented here
  and many other places

13
(No Transcript)