When we switch to use DFN-Mailsupport for spam and malware filtering, this results in some changes for mail delivery

How does this affect me?

• Spam und Malware is rejected.
  Messages that have been classified to be undesirable with relative high certainty ("Spam") will no longer be delivered, but rejected by the DFN. The sender will receive an error message. Mails with detected malware will also be rejected.
  Mails that are only probably spam will still be delivered. The established header lines (X-Spam-Status, X-Spam-Level) will then be added.
• "SPAM Hits=" is gone.
  We no longer tag subject lines with "SPAM Hits=". The new tag is "[Spam]" and no longer contains a spam score. If you configured your own spam filter based on the subject header, please adopt these and consider switching to evaluating the "X-Spam-Flag" header instead.
• Custom filter settings are no longer evaluated
  Up until now it was possible to customize the behaviour of the spam filter via the webmailer. This customization is no longer supported and any custom settings are ineffective after the switch.

How is the switch taking place?

We will switch over email domains one by one without individual announcement. We expect no negative effects on email delivery during the switch.

adaptations after the switch

Hermes Mailboxes

DFN uses a different threshold (6.2 instead of 8) for the spam score. If you use the standard ruleset for filtering, you might want to increase the sensitivity your ruleset.

Deletion rules from the standard ruleset are not longer needed. High-scoring spam is rejected via DFN.

If you have manually created filtering rules, you might need to adopt these. This is particularly required to rules that check for the "SPAM Hits=" subject tag.

Exchange Mailboxes

Exchange mailboxes use a central tagging mechanism. No changes should necessary for Exchange users.

If you have manual created filtering rules, you might need to adopt these. This is particularly required with rules that check for the "SPAM Hits=" subject tag.