On Jun 8, 2007, at 2:46 PM, Aaron Stone wrote:
On Fri, Jun 8, 2007, Michael Monnerie <[EMAIL PROTECTED]
management.at>
said:
On Freitag, 8. Juni 2007 Tom Allison wrote:
But is DCC going to want you sweeping 1000's of emails a minute =A0
against their servers to see if they're spam yet?
I meant you calculate the DCC checksum on arrival as usual, and
store it=20
in a field. Every hour or so, you make a GROUP BY to check how
often=20
each signature got hit. And then, if you've got some extreme
values=20
(e.g. 30% of all e-mail in the last hour had the same checksum),
you=20
check this one checksum again against DCC. There's a high
probability=20
they list it as spam already, so you can reprocess your e-mail
locally.=20
That way, it's only one extra lookup on DCC.
You're absolutely right that this information would have to go into
the
database. I think it would be possible to do it as add-on tables
with some
foreign keys into the dbmail schema, and that way we don't have to
build
it deeply into DBMail. If there's a really good case for building
it in,
then I'd certainly think about doing it.
I see what you are trying to do.
It should be noted that the effectiveness will be proportional to the
number of users/external email you process.
BTW -- how effective is DCC? Never used it myself.
_______________________________________________
DBmail mailing list
[email protected]
https://mailman.fastxs.nl/mailman/listinfo/dbmail
