Re: [moca] Locations?

[Martin Berger <martinb@xxxxxxxxxxxxxx>]

i agree: "Failures are a crucial point".

> What distinguishes a distributed system from a non-distributed one is that
> you can have parts of a computation failing (in various ways).

when tinkering with pi-based models of distributed computation, i could not
prove that location, location failure, message loss, message duplication or
checkpointing do in fact really extend the underlying non-located and non-
failing asynchronous calculus. i suspect they do, but maybe the failure
model was too benign for the separation to be drastic enough to allow a
simple proof. of course once one adds timing, separation is straightforward,
but that's already the case for non-distributed calculi.

what i wonder is if there's a more intrinsic notion of partial computation
than that given by explicitly introducing a location operator, one that can
already be defined in standard pi-calculi or the like.

> Calculi can only claim to be 'distributed' in my view if they allow
> you to model at least some form of location failures.

would you consider message loss between locations a form of location
failure?

martin


 
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
The "models for mobility" mailing list     mailto:moca@xxxxxxxxxxxxxxx
http://www-sop.inria.fr/mimosa/personnel/Davide.Sangiorgi/moca.html