As noted about a zillion messages ago, one of the concerns about IPv6 mail
is whether DNSBLs will be workable, with one of the questions being
whether the lookups will blow away DNS caches. As far as I can tell,
there is basically no research on DNS cache behavior other than a few very
old papers that said for normal loads, a reasonable sized cache is big
enough to work.
For the DNSBL stuff I'm doing some simulations, which need data to drive
the lookups. The data is really simple, a sequence of pairs:
[ IP address, timestamp ]
IPv4 traces are currently more useful than IPv6, since v6 doesn't have the
botnet traffic that makes the DNSBL lookups interesting.
Traces from medium sized systems, say a million connections a day, would
be great. If you consider the IP addresses confidential you can hash
them, although real ones would be better so I can test tweaks that treat
known high volume senders (gmail etc.) differently. Thanks.
Regards,
John Levine, jo...@iecc.com, Primary Perpetrator of "The Internet for Dummies",
Please consider the environment before reading this e-mail. http://jl.ly
PS: I hope it's OK if we take a break from reinventing patented anti-spam
techniques that have failed repeatedly before to talk about something
else.
