All numbers here do not take DNS TTL caching into account. These are the
raw numbers. It is believed that typical TTL caching of 24 hours would
reduce the usage amounts considerably.
Out of 409529 records over 18 months:
Total allows: 395800
Total Denys: 13729 (3%)
Bandwidth saved or gained:
Out of 19791 domains
Less than 100% bandwidth used: 380 ( 1%)
Up to 101% bandwidth used: 6605 (33%)
Up to 110% bandwidth used: 17881
So 33% of domains sampled here came up "awash" where no bandwidth was gained
or lost. I suspect this could've been a higher percentage if I knew more
about the .pl name space, as most of the entries I have were from Pik over
there.
The more popular domains I knew something about, and could obtain MX records
for, saved more bandwidth than those I didn't.
Domains that showed close to or just over 100% usage compared to normal
still means no forgeries originating from those domains. Is the extra
bandwidth worth not seeing that e-mail? I don't know.
Examples for popular domains:
domain msgsize dmpsize totalsize percentDiff
yahoo.com 104283339 1923494 84708036 81.23%
aol.com 98541811 1029857 89774170 91.10%
hotmail.com 119297701 1621395 94342602 79.08%
yahoo.com.tw101711 10792 55134 54.21%
returns.groups.yahoo.com
286849696 4948214 291797910 101.73%
DMP lookup bandwidth used compared to SMTP bandwidth:
Out of 19791 domains
Average additional bandwidth used: 4.15%
Less than 1% additional bandwidth: 6281 (31%)
Less than average additional 12635 (63%)
Greater than average additional 7136 (37%)
This is what happened regardless of wether a message was allowed or denied.
There was an average increase of 4.15% in bandwidth used compared to not
using DMP. Well over half of the domains sampled were at or below average.
A smaller sampling set this at about ten percent. Seeing a larger sampling
with a smaller percentage is encouraging.
It's interesting that almost a third of the sampled domains showed less than
one percent of an increase. I'm guessing the messages originating with
these domains as the sender were larger (HTML, rich text) or had large
attachments sent on a routine basis.
If anyone wants a copy of the Access 97 database I cooked these numbers
with, ask. I'll compact it and zip it for you.
--
PGP key (0x0AFA039E): <http://www.pan-am.ca/[EMAIL PROTECTED]>
What's a PGP Key? See <http://www.pan-am.ca/free.html>
GOD BLESS AMER, er, THE INTERNET. <http://vmyths.com/rant.cfm?id=401&page=4>
This is the discussion list for the IMS Free email server software.
To unsubscribe send mailto:[EMAIL PROTECTED]
Delivered by Rockliffe MailSite
http://www.rockliffe.com/mailsite
Rock Solid Software (tm)
