On Fri, 26 Nov 2004, David Woodhouse wrote:
It would certainly be interesting to see a study of how CPU load on mail
servers would be affected by DK; especially one which recognises the
fact that you don't actually have to run other CPU-intensive checks like
SpamAssassin and virus checking on mails which were already rejected due
to the crypto check.
If I remember right, two studies were presented at the FTC Summit. I took
some notes but looking at them right now, they were too brief and not in
good shape. So if somebody else who was there took better notes, please
correct what I got wrong. Anyway here is what I have:
1. coldspark.com [probably who presented the study]
Inbound Outbound
CPU Speed CPU Speed
IIM 55% -16.3% 55% -15.5%
DK 55% -5.4% 55% -5.5%
[Now those numbers don't quite look right to me (not sure what speed meant
and why dk is lower them iim),but I assume it means 55% increase in cpu
computation, not sure what key size]
Then on note sheet I see:
IIM: 864k/hour on single server
DK: 996k/hour (the same)
- 512 size key?
[I would assume this was on some kind of p4 dual xeon server simulating
sending signed emails with number showing how many emails it was able
to process. If I remember a question was asked what kind of emails
(size, type ) and answer was that it was fairly small about equivalent
to average email message and that this was test simulating average
email processing]
2. In separate box on the same note sheet I see:
512 bytes - 20% cpu higher
1024 bytes - 90% cpu higher
[I think this was DK with test on how much more cpu is needed
depending on size of the key]
-------------------------------------------------------------------
And getting back to SPF, I took some notes on the statistics that
were presented at well.
1. Godaddy statistics (?)
7% of emails go Godaddy have SPF records
18% of emails are rejected based on SPF
14% of SPF emails are from known spammers
2. Earthlink numbers
90% of emails that passes SPF is spam
90% of emails that fails SPF is spam
40% of emails that does not publish is spam
-------------------------------------------------------------------
On separate note sheet I have the following from the 1st day:
NIST person said that "SPF has largely been consumed by SenderID"
[and note to self to remind them that it is not]
Harry Katz: "I'll be bold enough to call it [SenderID] emerging standard"
"I certainly would not recommend outright rejection of the
message based solely on SenderID"
My notes also say that there was very bad guy from "Association for
Competitive Technology" on the first panel who tried to argue that
opensource should be ignored when accepting a standard (my notes
have such words for him that I can't even put in public mail list).
He also had long whitepaper on that topic named "Open Standards Patent
Policies and Open Source Software Implementations" that I picked up.
I'll scan and let you guys read it. On the paper their address is listed
as "1413 K ST, NW, 12th floor, Washington DC | www.actonline.org".
After seeing all that I understood that he was lobbyist paid by
"you know who" to say "you know what" and this association despite
its name is actually setup to promote anti-competitive standards
and technology. I'm warning you about it right now because you may
see this name come up in regards to SenderID or something else
that M$ is trying to push through.
--
William Leibzon
Elan Networks
william(_at_)elan(_dot_)net