More on Spamhaus ZEN
Yesterday I wrote about how Al Iverson had observed zero false positives from using the Spamhaus ZEN blocking list. I also noted that the sample set was too small (~2300 emails) to definitively give Spamhaus the Zero False Positives Seal of Approval.
Today, numbers provided by the Dutch ISP XS4All were brought to my attention.
In a nutshell, before ZEN, XS4All was using a combination of SBL + XBL + dynablock.njabl.org (ZEN is simply a combination of the SBL, the XBL and the PBL). With this combination, they were blocking about 4 million messages a day, out of 8 million examined. Their abuse@ address receives complaints about false positives about once every two weeks. Assuming that only one false positive in a thousand actually generates a complaint, that's a false positive rate of .00089%. Not too shabby.
(Of course, the one-complaint-in-a-thousand number is plucked straight from the air; I have no idea if anybody has ever done a study to find the actual number, but I think it's a reasonably conservative guess.)
The full article (in Dutch) can be found at Vincent Schönau's blog at XS4All.
Today, numbers provided by the Dutch ISP XS4All were brought to my attention.
In a nutshell, before ZEN, XS4All was using a combination of SBL + XBL + dynablock.njabl.org (ZEN is simply a combination of the SBL, the XBL and the PBL). With this combination, they were blocking about 4 million messages a day, out of 8 million examined. Their abuse@ address receives complaints about false positives about once every two weeks. Assuming that only one false positive in a thousand actually generates a complaint, that's a false positive rate of .00089%. Not too shabby.
(Of course, the one-complaint-in-a-thousand number is plucked straight from the air; I have no idea if anybody has ever done a study to find the actual number, but I think it's a reasonably conservative guess.)
The full article (in Dutch) can be found at Vincent Schönau's blog at XS4All.
0 Comments:
Post a Comment
<< Home