On Wed, 5 Mar 2003, Alan DeKok wrote:
Nah. If you're receiving that kind of mail volume, you have big
machines with under-utilized CPUs anyway. Content-scanning just uses up
CPU cycles that would be wasted. Nik Clayton says he filters
2 million messages/day with SpamAssassin; I'd be amazed if fewer
than 100K of those are spam.
With the spam volume Chris Lewis is saying Nortel gets, I'd be
surprised if their infrastructure costs were less than $100K.
Large companies can afford that. Small ones can't.
Small companies don't get 100K spams/day.network
Small companies routinely get several hundred, up to several thousand
spams/day.
I found the biggest improvement by blacklisting Korea and China. Mail from
IPs within their netblocks is simply refused. There are a few
South-American countries that I may add next. Using an RBL in-progress
spam attacks and another for open relays has reduced the spam that makes
it through from thousands to only a dozen or so per day.
Unfortunately, this still wastes my bandwidth. Server utilization isn't
really an issue, since even a slow box by today's standards can handle a
huge amount of rejected connections.
Blacklists (fixed and dynamic) are the technological equivilant of
"shunning", and seem to be quite effective if practiced on a widespread
basis. It can range from "If you break our social contract, we will ignore
you", to "we will ignore your network", or even "we will ignore your
region of the earth" if the spam continues.
The trick is to make spamming painful for the network owners, so THEY will
control the spammers. Networks are expensive and nobody wants to own one
that can't talk to anybody else.
At some point in the world of spam, there is a network operator who has
allowed a spammer a connection. This may be an ISP, Network Provider or
simply an open relay. It really doesn't matter. The place where the mail
enters the netwok should feel pain for polluting our space.
Once this happens, spam will again become only a minor annoyance.
Terry
_______________________________________________
Asrg mailing list
Asrg(_at_)ietf(_dot_)org
https://www1.ietf.org/mailman/listinfo/asrg