Hi,
This is embarrassing since the GRASP draft is already with the IESG,
but better now than later.
Bill Atwood has been testing the GRASP prototype on a larger setup
than previously, and this has just thrown up an issue in discovery
relaying. (We tried to do this test at the IETF98 hackathon, but
failed for practical reasons.)
The issue is minor in terms of words but really needs to be fixed.
Consider the following two extracts:
> 3.5.4.4. Discovery Relaying
...
> Since the relay device is unaware of the timeout set by the original
> initiator it SHOULD set a timeout at least equal to GRASP_DEF_TIMEOUT
> milliseconds.
> 3.8.5. Discovery Response Message
...
> It MUST contain a time-to-live (ttl) for the validity of the
> response, given as a positive integer value in milliseconds. Zero
> is treated as the default value GRASP_DEF_TIMEOUT (Section 3.6).
This is exactly the wrong way round. The TTL for a discovery response
needs to be *longer* than the timeout for a (relayed) discovery. If not,
the TTL will expire before the relayed discovery completes. This is
really an implementation issue but the above text almost guarantees
failure.
Proposed new text, which intentionally leaves the details for the
implementer:
> Since the relay device is unaware of the timeout set by the original
> initiator it SHOULD set a timeout significantly less than GRASP_DEF_TIMEOUT
> milliseconds.
...
> It MUST contain a time-to-live (ttl) for the validity of the
> response, given as a positive integer value in milliseconds. Zero
> implies a value significantly greater than GRASP_DEF_TIMEOUT
> milliseconds (Section 3.6).
WG Chairs and AD: I will be on personal travel between now and the
IESG call on Thursday, so it isn't really possible to post a new
draft in time.
Regards
Brian
_______________________________________________
Anima mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/anima
