> After a couple of tests, it looks like all I have to do is check the
> From  field...

Yep.

> ...parse the X-RCPT-TO line...

No,  you  won't have the X-RCPT-TO in files that are delivered to PAs.
The way you determine the RCPT TO: is by parsing out a preset argument
from  the  PA  command  line.  Just  as with lists, which are PAs, you
create  a  PA for each distinct user (same exe, different args), since
you  can't  ensure  that  the  header/body  will  have either piece of
envelope information.

> ...and append the email to the *.mbx file.

Uh,  no.  DO  NOT  do this. Send passing mail with IMAIL1 or inject it
into  the  queue and call SMTP32. Otherwise, the index files will have
to be rebuilt for EVERY message: you want overhead, you got it!

> The advantages to this seem to outweigh the other options.  After much
> discussion, it's also been determined that outgoing filtering on a per user
> basis is a "must have".  Since there is only 1 orules.ima file for the whole
> domain, it would be impractical to rely solely on the one orules.ima file
> for outgoing mail of 5,000+ users.

Outgoing  mail  is  not going to go through a program alias. You would
have  to  write a full-fledged SendName for that purpose. And it's not
simple  (as  I  often say, Scott invested plenty of time to understand
the   interoperability  process,  even  before  Declude's  proprietary
functionality was added).

> Cussing in an email will get the same treatment as cussing in class.

Wow, time for a second Hotmail account. :)

> Our only concern with the program alias is the processing overhead.  Is
> there a hard limit to the number of simultaneous program alias spawned
> processes?

No,  but  there  is a practical limit to the number of service-started
processes  under  Win32:  see  Desktop Heap issues, the same ones that
plague  IMail  in general and which Declude treats. You may, depending
on how long your exe stays in memory, have to treat this issue.

>  If  we break the 10,000 user mark (which will come in Spring 2003),
> will message delivery speed suffer significantly? I guess we'll just
> have to see.

Yes,  as  is  often said, it's your usage, not your user count, that's
the  relevant  statistic.  And  again,  in  this  case,  the  ratio of
failed-to-passed   traffic  will  have  a  direct  impact  on  overall
throughput;  as people learn to play by the rules, there will be fewer
wasted cycles.

-Sandy


To Unsubscribe: http://www.ipswitch.com/support/mailing-lists.html
List Archive: http://www.mail-archive.com/imail_forum%40list.ipswitch.com/
Knowledge Base/FAQ: http://www.ipswitch.com/support/IMail/

Reply via email to