Skip to content

DMARC Implementer Notes

Matt Simerson edited this page Jun 18, 2014 · 9 revisions

A DMARC implementer might decide that a good data model to use is the XSD schema included in the DMARC spec. If one validates against that specification, they should be aware of the following data errors observed in the wild:

XSD schema deviation

Missing record section

From June 06, 2013 until ..., Microsoft has been sending out DMARC reports with missing record sections.

Mis-labeled XML tags

Reports from junc.org have mis-labeled the 'identifiers' section as 'identities'

Invalid SPF result

Reports from ivenue.com arrive with /auth_results/spf/result=unknown. Unknown is not a valid SPF result.

Gzip -vs- Zip compression

The 2012 and prior drafts specified zip compression. The 2013 draft specifies gzip compression. More details are available in this mailing list thread. Mail::DMARC receives reports in both formats and sends in gzip format. Mail::DMARC may be the only implementation to send in gzip format. There are at least two other implementations that read gzip compressed reports:

Date: June 4, 2013 6:12AM
Subject: DMARC support for ingesting gzip'd files

Thanks [email protected]!

We just rolled out importing of gzip'd DMARC XML, as you're the first
to send files this way.  Everyone else has been sending via ZIP (not gzip)!

Well done!
The dmarcian.com crew

As of June 10th, I have yet to receive a gzip report.