I've been using them for years. We do a lot of email (5 mail servers) as
an ISP.
I sometimes get network test access for free, for others I have paid.
It's either a pay big or pay nothing, with no middle ground
unfortunately. Many of these, I run my own dns servers and use rsync to
replicate their data on virtual machines separate from my mailservers.
There is a big investment in bandwidth to establist a blocklist if you
want it to be useful and popular. There is also a technical hurdle to
the people running the mail servers to utilize systems for rsyncing,
running third party nameserver software, bind integration, etc... It's
very rigid, hierarchiel, manual patchwork. Most of these systems are not
using any sort of secure DNS either.
I wouldn't mind running a DNS or rsync server for some of these lists,
to give back to the community, but everyone seems to do it differently
or have different qualifications or credit for such tasks. It's just
easier to not even suggest it.
I suspect network tests are a popular and growing portion of the future
of antispam. I'm not a coder, I'm a CS dropout, so I can't do this
myself. It's more of a vision than a concrete software product, so hear
me out and discuss this.
Integrate some sort of private purpose p2p system with spamassassin. All
the *bl network test managers could constantly "seed" the p2p system
with up to date data as they create it. The data could have a lifetime
like DNS data does. The SA p2p module/daemon would gather the data for
the *bl systems one is subscribed to and make it available for local or
local network spamd tests. It could also make it available to other SA
p2p system via the p2p network.
Thus someone could have a major and popular *bl and not have to have the
huge bandwidth and management needs some of the bigger semi-commercial
blocklists certainly have.
A large user or ISP with excellent bandwidth and servers could also be
contributing bandwidth and server capacity to the SA p2p *bl system with
minimal technical hoops. They could also scale big without causing any
load issues for the creators of the *bls.
SA-updates could also be distributed via the system.
As an ISP, I've generally been wary of p2p as it is so inefficient to
distribute data from the far reaches of the network to the other far
reaches of the network instead of a layout focused on a high capacity
core. However for such a project, it would likely be primarily well
connected servers, and p2p system could be efficient and practical.
--
/*
Jason Philbrook | Midcoast Internet Solutions - Wireless and DSL
KB1IOJ | Broadband Internet Access, Dialup, and Hosting
http://f64.nu/ | for Midcoast Maine http://www.midcoast.com/
*/
