Synchronous modules in ad hoc distributed embedded real-time system ( AdhocEsterel)

Integrating the synchronous formalism into the complex,  asynchronous, real environment
was a work plan in the "Synchronous Reactive Formalisms"  (SYRF, 1997-99) project.
In the "Advanced design tools for aircraft systems and airborne software" (IST SafeAir, 2000-02)
project the data flow synchronous language, Signal, with its per component activation clock
(multiclock or polychronous model) was the basis of a platform called Polychrony.
Polychrony  was demonstrated with avionics software on  the Integrated Modular Avionics
architecture. However, no such success is seen with Esterel, the imperative synchronous
language which is a better design and implementation choice for control oriented applications.
This lacuna is probably due to the single-rate synchronous concurrency model of classical Esterel.
In the IEEE standardization effort for Esterel V7, multi-clock models are recommended. Further,
there are efforts in Germany and New Zealand to target Reactive multi-processors for Esterel.

Real systems are not only distributed and asynchronous but also open. An ad hoc Esterel module
may  have to be  added  to  solve a problem that crops up. The project objective is to study the
requirements and constraints on the compiler, run time executive and ad hoc modules of Esterel,
create a tools platform (similar to  Polychrony for Signal) and  demonstrate the  suitability  of the
platform with a real world application (similar to the avionics for Signal).

Scope

Often safety-critical systems cannot be designed, operated or even started as a monolithic rigid
closed  system. The technologies available for compilation and execution of Esterel should not
limit its use to monolithic systems. The compilers and run time excutive for the synchronous
imperative language Esterel must support, not only distributed but also, ad hoc architectures.

Necessary steps

1. Extend an Esterel compiler (similar to the EEC2 compiler by Univ of Aukland, NZ), to produce
    modular target code for a closed set of Esterel modules.
2. Develop the Run Time Executive that runs the target code modules  in conformance with Esterel
    concurrency and synchrony models. The Run Time Executive and the stock processor should
    look like a Reactive Esterel Virtual Machine.
3. Explore constraints on the specification and execution  ad hoc Esterel modules.
4. Modify the compiler of step 1 to produce modular target code for ad hoc Esterel modules.
5. Extend the Run Time Executive of step 2 to  execute target code of ad hoc Esterel modules.
6. Demonstrate the approach and tools with a real world, significant application like avionics.

Results

  After step 2: Compiler and Run time executive for modular target code from monolithic
                       Esterel program.
  After step 3: Characterization of ad hoc Esterel modules.
  After step 5: Compiler and Run time executive for ad hoc Esterel modules.
  

synchronous programming, ad hoc module, distributed system, Esterel, Esterel Virtual Machine, polychronous, multi-clock

Experts in synchronous programming,  Esterel compiler, distributed embedded  safety-critical and
reactive real-time  system,  reactive processors. Project coordinaton.

Theoretically underpin distributed ad hoc Esterel. Academic researchers.

Modify and extend Esterel compilers and run-time.  Academic or industrial researchers.

Incorporate distributed ad hoc Esterel formalism into development suites. Industrial researchers.

Apply the distributed ad hoc Esterel formalism and development suite with an industrial scale
application such as avionics.   Industrial researchers.

Manage the project.  Industrial partner with prior experience as  Coordinator of similar projects.

Industrial partners in Safety critical systems  (avionics, transportation, medical and nuclear power)

University and academic researchers in Synchronous reactive formalisms.

Project coordinators with experience in safety critical or reactive systems.

Yes