ietf-dkim
[Top] [All Lists]

Re: [ietf-dkim] Updated implementation report

2010-09-29 18:12:26


-----Original Message-----
From: ietf-dkim-bounces(_at_)mipassoc(_dot_)org [mailto:ietf-dkim-
bounces(_at_)mipassoc(_dot_)org] On Behalf Of Hector Santos
Sent: Wednesday, September 29, 2010 5:23 PM
To: ietf-dkim(_at_)mipassoc(_dot_)org
Cc: ietf-dkim(_at_)mipassoc(_dot_)org
Subject: Re: [ietf-dkim] Updated implementation report

Murray S. Kucherawy wrote:
I've posted a new issue of the DKIM implementation report.  The most
interesting changes are the inclusion of a day of sample data from AOL
and
a revision of the data summary reported by the OpenDKIM stats project
using the updated schema, which allows a few interesting new
observations.


Outstanding job Murray. Is this sufficient to achieve the intended goal?

Thanks for the report.

I believe the most outstanding data point here is the extremely high
original signing of mail.  AOL's 1.2 out of 1.4 billion or 86%
signatures were 1st party and your own statistics showed 73% were 1st
party.


This isn't particularly surprising to me. I've always believed there is
a significantly higher value for 1st party signing compared to 3rd
party. The value proposition has always seemed to be much more
compelling to me.

It would be interesting if the data could be sliced and diced a little
differently. 

For example, it would be interesting to look at percentage based on
volume compared to percentage based on number of domains. In other
words, are a small number of large domains skewing the reported numbers.

It would also be interesting to look at breakout percentage of domains
signing based on number of emails received from that domain. So,
arbitrarily segment the set into quintiles based on number of mails
received for that domain and then look at signing practices within the
quintile.

Mike

_______________________________________________
NOTE WELL: This list operates according to 
http://mipassoc.org/dkim/ietf-list-rules.html

<Prev in Thread] Current Thread [Next in Thread>