Would anyone be interested in receiving a text or BGP feed of IPs of hosts known/suspected to be compromised and used as parts of DDOS attacks? Would anyone be interested in contributing their BGP views?
We have (and I'm sure we're not isolated) been seeing attacks from several thousand/tens of thousands of unique hosts generated >2Gb/s, >1Mpps attacks.
I am not necessarily suggesting that providers use this list to blackhole at their edge, but its certainly a good candidate for that. It could alternatively be used by access providers to notify their customers or filter on their customers. I am sure it would also be a good list to use to deny traffic to SMTP servers from/to.
I'm not really an activist, so if there is real interest, I will be glad to set it up and contribute our own significant list of sources.
If this is already done and I don't have a good set of skills with Google, please let me know.
We're doing this internally, watching for various types of attack probes (SQL Slammer, Mydoom, dictionary attacks over SMTP, Nimda, etc.) and lock out source addresses via BGP blackholing for those who are persistent. All blocks age out over time so that systems that get fixed are removed by virtue of the attacks stopping. At any given time we have blocks against 800 to 2000 systems.
At present we don't make this available to anyone outside, though it wouldn't be that hard to do.
It also has nothing to do with DDoS attacks; it's intended use is only
for blocking email traffic. The XBL incorporates the CBL, and the CBL
team does not support the use of its data for purposes other than
blocking incoming SMTP traffic. The reason for this is that the CBL
lists a very large number of dynamic IPs, and has a very long
expiration time (months). Accordingly, using it to block general
traffic will have a high false-positive rate.