A Review on: Strategic Directions In Real Time and Embedded Systems: John A Stankovic Et AL'

1
A Review onStrategic Directions In Real Time
and Embedded Systems John A Stankovic Et AL.

• By Aditya ShuklaDate 09/19/2006

2
Agenda

• Summary of Paper.
• Strength of the approach.
• Weakness of the approach.
• Relevance to Embedded Systems.
• Conclusions
• Short Video on Real Time Application.

3
1.Summary of Paper

• 1.1 Real Time Computing ????
• 1.2 Real Time Computing Applications
• 1.3 Growth Related Statistics
• 1.4 Future Challenges
• 1.5 Future Vision
• 1.6 Summary

4
1.1Real Time Computing??

• 1.1.1 Correctness of results are not exclusively
  based on logical results, they also rely upon the
  time at which results are produced.
• we may explain Real Time system graphically

5
1.1.1 continued..

• A Non - Real Time System No Time Axis.

Result F(X,Y)
Y
6
1.1.1 Continued..
Y
Result F(X,Y,T)
(X , Y, T )
X
T
7
1.1.1 Continued..

• Mostly real time systems are the components of
  larger systems. (i.e. they are embedded systems)
• Real Time System A Classification. 1. Soft
  Real Time Systems -Latency range is
  accepted. -System failure doesnt lead
  catastrophes. (Examples Audio-Video Streaming,
  OLAP uses in Data warehousing.)
• 2. Hard Real Time Systems (or Safety
  Critical Real Time Systems)
• -Vary Low or Practically no
  Latency expected.
• - System failure leads
  fiasco.(Examples Health Monitoring Systems,
  OLTP uses in Data warehousing.)

8
1.2 Real Time Computing Areas Applications.

• 1.2.1 Real Time Computing DomainsAn enabling
  technology for many important application areas.
  E.g. Process Control Nuclear Power
  Plants Agile Manufacturing
  Intelligent Vehicle Highway System Avionics
  Air-traffic Control Telecommunication
  Multimedia Applications Medical
  Applications Defense Systems All
  Safety Critical Systems ..and
  many more..

9
1.2 Real Time Computing Areas Applications.
Contd

• 1.2.2 Real Time Applications.
• 1. Air Traffic Control System
• 2. Digital Control and Constant bit rate
  Audio Video. 3. Real Time Computing in
  Business.
• 4. Automobile Industry Implementations
• 5. NAVSTAR GPS system.

10
1.2.2 Real Time Applications. Contd

• 1.2.3 All Applications carry in common -
  Precision in Mili-Seconds
• - Latency is minimal - System failure may
  lead catastrophe.

11
1.3 Growth Related Stats

• Over two billion dollars are spent annually on
  tools, application S/W and embedded OS
• Commercial real time O/S market has reached 100M
• Many mission critical systems have adopted real
  time OS. (Ex. Boeing, McDonnel-Douglas.)

12
1.4 Future Challenges

• 1.4.1 High Level
• System Evolution
• Open Relative System
• Composibility
• S/W Engineering

13
1.4.1 Contd

• System Evolution
• Important Features of this new real-time software
  architecture.
• Extensive use of open Standard Systems
• Implementation of Coherent set of interfaces for
  the integration process control, plant wide
  scheduling and plant management information
  systems.
• A convenient and safe environment for
  customization, optimization and configuration of
  systems.

14
1.4.1 System Evolution Contd.

• Trouble free replacement of obsolete systems
• For COTS-
• Developing scheduling and resource management
  schemes for systems for demonstrable predictable
  properties.
• Composing subsystems with predictability into
  bigger systems.
• Making subsystem essentially re-usable.

15
1.4.1 Contd

• Open Real Time Systems
• Build and deliver general purpose open real time
  system.
• Developing a dynamic mix of multiple and
  independent real time applications which can
  coexist on same machine and set of machines.
• Incorporating real time and non-real time systems
  together.

16
1.4.1 Contd

• Open Real time systems Challenges
• Hardware Characteristics are unknown until
  runtime.
• The mix of applications and their aggregate
  recourse and timing are unknown until runtime.
• Perfect priori schedulability analysis is almost
  impossible.

17
1.4.1 Contd

• Composibility
• Real time should be robust while delivering HIGH
  real time performance. They need to evolve and
  use legacy components.
• Developing high performance fault tolerant
  embedded systems which take care of Interacting
  domains, function time and fault tolerance.

18
1.4.1 Contd

• S/W Engineering
• Time, dependability and other QoS constraints
  must become first class.
• Platform independence concern should be separated
  from application level concerns.
• Software must be structured into composible
  modules.
• Software must be designed to be adaptive and
  configurable.
• Timing constraints should be derived component
  wise and must be derived dynamically.

19
1.4.2 Low Level (Challenges)

• Science of Performance Guarantees
• Reliability and formal Verification.
• General System Issues
• Real time multimedia and Programming Languages.
• Education

20
1.4.2 Low Level (Challenges) contd

• Science of Performance GuaranteesThe collection
  of algorithms and analysis that exists for
  static, deterministic guarantees.-Extend for Real
  Time Systems
• System becoming more larger, more dynamic and
  deployed in non-deterministic-we need to expand
  SPG.
• Timing validation and workload are key areas for
  SPG research.

21
1.4.2 Low Level (Challenges) contd

• Reliability and formal Verification
• Better Understanding of checking correctness of
  real time systems at runtime.
• Using design time model checking
• Using design time model checking and simulation

22
1.4.2 Low Level (Challenges) contd

• General System Issues
• Architecture implementation and changes needed
  for worst case execution times?
• Real time muticasting be sufficiently
  implemented??
• Effect of various kind of caches??
• H/W and S/W co-design be done for real time
  systems?
• Good architecture and protocols for real time
  systems?

23
1.4.2 Low Level (Challenges) contd

• Real time multimedia and Programming Languages.
• Application level predictability requirements
• QoS requirements.
• Real time queuing algorithms requirements
• Requires a language which can manage a wide range
  of requirements, timing guarantees and scheduling
  approaches via meta-level programming.

24
1.5 Future Vision

• All products and process will have some kind of
  real time features.
• We will have more and more open systems in
  future.
• Real time systems will be achieve high degree of
  precision hence cater a better QoS.

25
2.1 Strength of Approach

• A detailed discussion with examples for past,
  current and future real time systems.
• Covered almost all requirements for highly
  dynamic real time systems.
• A clean and witty approach to envision future
  real time systems.

26
3.1 Weakness of Approach

• Details are given at very high granularity level.
• Major emphasis on logical level of Real Time
  computing, most of the time system level and H/W
  constraints are not discussed in detail.

27
Relevance to Embedded System

• Most of the real time (sub)Systems are embedded
  ones.
• Examples 1. Early Warning Aircraft Systems
  2. Automobile Fuel injection system 3. Air
  traffic control system 4. NAVSTAR GPS system.

28
An Example