I am new to POE (been at Perl for a LONG time
but I still learn things about it daily) and
becoming quite a fan.

My mail server can check "Realtime Blackhole
List" DNS servers and checking one or two has
reduced our SPAM by 90%.

Currently, I am working on an RBL-DNS Server
service which will run initially on Win32 to
give me the ability to check more RBLs, in
parallel, whitelist/auto blacklist (without
checking on the net), combine multiple RBL
reports (additive scoring), and cache the
results to improve performance.

Receive DNS request in POE, check cache and
return immediate answer, or dispatch event to
query MULTIPLE RBL servers, receive answer and
immediately respond on 'success' or after all
(or percentage of) queries return NOT FOUND,
cache results.

Note:  SpamAssassin does much of this and I am
also working on running it on Win32 with my email
server but SA checks AFTER the email is received
under my current architecture but the RBL check
immediately DENIES the connection before email is
spooled.

(There are other architectures, checking through
SA first, then the email server, but even then
the RBLs are the biggest single successful SPAM
filter with a very LOW false positive.)

Are any of you working one something similar or
interested in this idea?

Using POE has taken me a week or so of casual
work, but as I get more comfortable with it this
is getting easier -- it also seems less necessary
as my understanding of the problem and multitasking
the solution is improving along with my knowledge
of POE and using it to solve this problem.

Anyone else notice this effect?  (i.e., a hard
multiple agent problem seems easier to handle
with 'ordinary' code after you work through a POE
solution)?

Herb Martin
[EMAIL PROTECTED] http://LearnQuick.Com
512 388 7339   -or-   1 800 MCSE PRO





Reply via email to