[SystemSafety] A question about ISO26262

[Martyn Thomas]

David,

I'll use the standard's term "element" as a generic word for functions,
processes and other bits of software, as I don't have the standard
available and it's not necessary to be more precise to make the point
that I want to make.

It's certainly possible for an element to meet its functional
specification whilst preventing another element from doing so - for
example by taking too many processor cycles, or holding a shared
resource, or by using storage that is also being used by another element.

Drew has expressed his view that an element that "spams" another has
failed, and I wouldn't disagree. But the practical question will often
focus on the assurance that a safety-critical function cannot be
disrupted by other elements, and it will generally be impractical to
assess the total possible behaviour of every system element in order to
assure the safety critical elements. In practice it will be necessary to
show that the safety-critical functions cannot be disrupted by any
elements at a lower assurance level (DAL, SIL or whatever) and this will
in general require architectural protection (e.g. separate processors or
a high integrity scheduler and supervisor). I think the UK CAA standard
SW01 may have addressed this in some detail (if not, I'm fairly sure
there's a paper by Adelard on the subject, possibly written by Robin
Bloomfield).

Regards

Martyn


On 16/02/2017 10:37, David Haworth wrote:
> However, it is possible to imagine scenarios where, in the absence
> of a protection mechanism, an element causes another element to fail
> while still continuing to perform its own function flawlessly.

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.techfak.uni-bielefeld.de/mailman/private/systemsafety/attachments/20170216/78efed62/attachment.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 560 bytes
Desc: OpenPGP digital signature
URL: <https://lists.techfak.uni-bielefeld.de/mailman/private/systemsafety/attachments/20170216/78efed62/attachment.pgp>