Over the last 3 months the amount of email handled by our mail system has increased by 720%.

More than 98% of this mail is unsolicited commercial email (spam). This is placing added preassure on our systems and causing our spam protection systems to become less effective.

One of the major contributing factors to this problem is the use of catch-all email accounts. This is where one email address will accept mail for anything sent to that domain.

In our continued efforts to protect our mail servers from both spam, virus and harvesting attacks we have implemented blocking of dynamic IP ranges.

This means that if the IP connecting to our mail servers returns as from a dynamic IP range we will drop the connection and not accept the mail.

After some recent reports by users I have looked in to the effectiveness of spam filtering on catch-all accounts. Internal testing and discussion with the mail server developers has brought out the fact that spam filtering is not applied to catch-all accounts.

Well that is not completely accurate. I should say spam filtering is not applied to catch-all accounts where the mail is received indirectly. Where mail is sent directly to the catch-all account spam filtering is applied as normal.

Example

With the recent increase in spam containing only images we have adjusted the scoring on emails containing a low text to image ratio. This should hopefully weed out the emails that are image only.