Microsoft already does what is in the spec in Office 365 (which they 
specifically call out in the draft), so eventually Hotmail/Outlook.com will 
inherit it. But there are two differences:

1. Office 365 stamps "dmarc=bestguesspass", not "dmarc=pass". That makes it 
easier to distinguish in the logs
2. We do relaxed alignment, not strict alignment like it says in the spec

Seems to work just fine.

Also, not sure why there would be a discussion of rua and ruf. It's 
straightforward to interpolate the virtual DMARC record but how can you 
interpolate where to send a failure report?

--Terry

-----Original Message-----
From: dmarc [mailto:[email protected]] On Behalf Of Kouji Okada
Sent: Friday, March 24, 2017 2:19 AM
To: dmarc <[email protected]>
Cc: Kouji Okada <[email protected]>
Subject: [dmarc-ietf] updating draft-akagiri-dmarc-virtual-verification

Folks

We are now working on revising draft-akagiri-dmarc-virtual-verification.
https://na01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fdatatracker.ietf.org%2Fdoc%2Fhtml%2Fdraft-akagiri-dmarc-virtual-verification&data=02%7C01%7Ctzink%40microsoft.com%7C75cd5739368a40ce682208d47296da95%7C72f988bf86f141af91ab2d7cd011db47%7C1%7C0%7C636259439620932357&sdata=lDAi6TjldCXogGZlQI1VpLfYOya3fjaJPRn8mtBgo1U%3D&reserved=0

We are going to submit new version in early April.
Authors are recognizing there are some open issues for the draft.

1. The authentication code.

In the draft, we suggest to mark “dmarc=pass” in the authentication result when 
the virtual dmarc verifications have passed.
We are going to keep it as it is.

In 02, we are going to mention that e-mail operators can add comments to the 
authentication result field to indicate the “pass” is marked by the virtual 
verification.
We are not going to define the format of the comments, but the example comment 
would be like below.

ex) dmarc=pass(vdmarc=true)

2. vdmarc=fail problem

When submitting 00 version of the draft, some people gave us the comments that 
it is harmful to mark “dmarc=fail” without explicit declaration of policy 
records.
Our intention is to utilize the authentication results of “dmarc=pass”
for the e-mails potentially can be treated as so.

As Takehito posted to this ML,
our evaluation on Japanese ISPs showed
more than 20% of received email traffic can be treated as "dmarc=pass" with our 
verification.
Thus our proposal helps to increase the number of legitimate e-mails on the 
receiver side.

“Statistics about effects of “Virtual DMARC””:
  
https://na01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fwww.vdmarc.dmarc.jp%2F%3Fp%3D122&data=02%7C01%7Ctzink%40microsoft.com%7C75cd5739368a40ce682208d47296da95%7C72f988bf86f141af91ab2d7cd011db47%7C1%7C0%7C636259439620942364&sdata=mVyeZ5EMI6cj1pvUXVYinZZi64JLqjE9v90iCUwiJ4M%3D&reserved=0

We are going to emphasize the point in 02.

3. rua, ruf

We do not define any default report destinations for the virtual verification.

4. Draft tile

Currently our draft title is “DMARC verification without record 
definitions(dmarc-virtual-verification)”.
Would you have any suggestions for the title?


Any comments will be appreciated.

Thank you,
Kouji Okada


_______________________________________________
dmarc mailing list
[email protected]
https://na01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fwww.ietf.org%2Fmailman%2Flistinfo%2Fdmarc&data=02%7C01%7Ctzink%40microsoft.com%7C75cd5739368a40ce682208d47296da95%7C72f988bf86f141af91ab2d7cd011db47%7C1%7C0%7C636259439620942364&sdata=AeIU1ls97f%2FktoX0ZuTufv1xDE0Q8%2FTAq%2BGpK8g9MvE%3D&reserved=0
_______________________________________________
dmarc mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/dmarc

Reply via email to