Brad Knowles wrote:
At 3:59 PM -0400 2003/08/13, Chris Lewis wrote:
=head2 Listings should be temporary
Listings should all be temporary (suggested default listing period:
24 hours L</Footnote1>) so that if your blacklist doesn't get around
to removing the entry then it times out at some point in the future.
Fine, but if the entry is removed and then you discover reason to
put it back again (maybe because of a re-scan, or whatever), then the
next timeout should be longer.
I'd recommend something like a bounded exponential backoff.
A bit further down:
=head2 Removals must be prompt
The preferred way of doing this is I<removal without question>. This
allows people to ask and get their IP address removed immediately, but
does not prevent the blacklist owner re-listing their IP address at a
later time (subject to seeing more spam, or re-checking the IP address
security). A re-listing due to a false removal request may result in a
longer timeout (i.e. a penalty).
Covers that. Agree tho, should mention "bounded exponential backoff".
I've fudged the original a bit.
Moreover, I think I'd probably move someone from the "confirmed"
list to the "probationary" list immediately, then investigate their
claims of being cleaned up. They might have to stay on the probationary
list for a while before I'd be willing to let them off completely. If
they ever came back, they'd stay on the main list a lot longer, have to
stay on the probationary list a lot longer the next time they claim to
be cleaned-up, etc...
We prefer an instant removal and relist upon subsequent re-detection or
re-test. Tho, a delay algorithm on removal (initially zero) is reasonable.
Other than that, it seems to be reasonable. I look forward to
seeing an expanded version that details tools and methods that can be
used to run a real-world black list.
I don't think BCPs should get too specific about methods/tools -
especially because they're often obsoleted quickly. It's important to
get the principles laid out.
_______________________________________________
Asrg mailing list
Asrg(_at_)ietf(_dot_)org
https://www1.ietf.org/mailman/listinfo/asrg