The big players have enforced rules against high-volume senders, but low-volume senders have been left alone. This is the first known application of rejects against low-volume senders.
Your suggestion merges two issues, spoofing and DMARC. Spoofing may be occurring if the DMARC test produces a result other than Pass, but you can only detect that possibility if the DMARC test is performed.
The DMARC test can be applied to any message, using the public suffix list and a default policy of relaxed authentication. The problem is that essentially everyone, including SmarterMail, has concluded that one must not perform the DMARC test unless the domain owner has given you permission to do so via a DMARC policy. This is foolish. Nobody has authority to tell me to ignore important information. When the test is omitted, many messages that are verifiably authentic are instead classified as "possibly inauthentic due to no data."
On the flip side, when a message does not produce DMARC Pass, it may still be acceptable. If you are going to enforce authentication, you need a data structure to provide alternate authentication when an acceptable message arrives without authentication by algorithm. For example:
Treat as DMARC Pass when:
- Message-From-Domain='seaworld.com' and
- SMTP-From-Domain = 'bounce.m.seaworldparks.com' and
- SPF-Result="Pass".
The same type of structure is needed for local authentication of the SMTP-From-Domain. For example:
Treat as SPF Pass when:
- SPF-From-Domain = 'example.com', and
- Server host name ends like '%.google.com', and
- Server host name is verified to the Source IP by forward DNS.
When these two structures are built, it is possible to enforce mandatory authentication. Messages that do not authenticate get routed to Quarantine. Acceptable message sources are given alternate authentication and unacceptable messages are block listed. Either way, the quarantine only occurs once, so the quarantine volume converges quickly to a minimal level.
Benefits:
- Spoofed identity never gets released to the user.
- Acceptable-but-unauthenticated messages are only quarantined once.
- Whitelisting is applied safely because identity is not spoofed.
- Content filtering could become user-specific because identity is not spoofed.
- Regular correspondents can be distinguished from new senders because identity is not spoofed. (Essentially all spam comes from new senders.)
Without those structures and an efficient quarantine review, enforcement of authentication will produce a steady stream of unwanted blocks.