Discuss July 2008

discuss@lists.surbl.org

9 participants
4 discussions

URI BL responses, always loopback network?
by Ron Guerin 05 Aug '08

05 Aug '08

Hello again, I posted here once before I believe on the subject of URL shorteners. I am the operator author of the GPL software known as TightURL, and operator of the .com of the same name where there is a public installation of this software. I was supposed to come back with some kind of response about checking URL submissions at submission time vs. re-checking later, but various things in life intervened, and I have yet to implement anything to track how many URLs I reject at submission time, but I am pleased to report that re-checking previously accepted URLs has clearly been a success. I am presently rechecking all my URLs every 6 hours *if* they meet the following conditions: * Is not already blocked in my database * Was added within a window period (7 days) -or- * Has recent activity and more than a threshold of hits recorded Using this formula, I've been blocking URLs that were submitted over a year prior to blocking them, along with many recent submissions. I have not added the capability to unblock them later, largely because visual inspection of the blocked URLs gives me a pretty good feeling they'll be blocked again sooner or later. I may have to do something about that, but I'm not seeing the need right now. Anyway,... today's question pertains to people who are mucking around in their DNS resolution, or have DNS providers that are falsifying DNS responses in order to generate ad income from typing errors in people's browsers. I recently concluded some support for a new user whose every URL submission appeared to be listed in both SURBL and URIBL, which my code checks to see if the URL should be rejected. This user was somehow getting his own IP address returned to him any time the true response was really NXDOMAIN. This was some local thing on his end, but I presume the same problem exists if your DNS provider is tampering with the responses, except in that case you'd get some IP address from the public IP space that misdirects you to some "helpful" site if you happen to be using a web browser. My workaround for this problem was to check to see if the first 4 characters of the response are "127." Every BL or URI BL I know about returns a response somewhere in the range of 127.0.0.0 - 127.0.0.255, unless I'm mistaken. While I guess this still leaves open the possibility for false-positives on my users end if they make use of both loopback network hosts and their DNS responses have been tampered with, what I wanted to know was if anyone knows of a BL or URI BL that returns or plans on returning something other than a loopback address as a positive response to a lookup? - Ron

3 5

Re: [SURBL-Discuss] RFC: significantly reducing FPs on OB
by Jeff Chan 20 Jul '08

20 Jul '08

On Sunday, July 13, 2008, 9:12:25 AM, Joseph Brennan wrote: > Jeff Chan <jeffc(a)surbl.org> wrote: >> I think we >> probably can't reveal the exact listing criteria in case they're >> useful for the bad guys. I know it's somewhat inappropriate to >> ask for comments without revealing details. I suppose I'm asking >> for general responses then. :) > So you'll keep ob, but take some undisclosed action to improve its > accuracy. Sounds worthwhile to me. Thanks! Yes, we would not get rid of OB entirely ever. It does have some good data, but with too many FPs. The goal would be to keep as much of the good data as possible while eliminating most of the bad. Unfortunately some of the good data may be thrown out with the bad; baby with the bathwater, so to speak. IMO FPs are much worse than FNs, so some increase in FNs balances out a decrease in FPs. Trying to decide if it's worth doing.... Jeff C.

5 7

Re: [SURBL-Discuss] RFC: significantly reducing FPs on OB
by Jeff Chan 13 Jul '08

13 Jul '08

On Saturday, July 12, 2008, 7:00:29 PM, Joseph Brennan wrote: > --On Saturday, July 12, 2008 1:41 AM -0700 Jeff Chan <jeffc(a)surbl.org> > wrote: >> We can probably significantly reduce the false positives on >> ob.surbl.org, the SURBL list based on Outblaze's URI blacklist: >> >> http://www.surbl.org/lists.html#ob >> >> at the cost of some possibly minor false negatives. [...] >> Should we make that change? > This agrees with our experience using SURBL for a few years. We've > seen the occasional fp with the ob listings, and none I am aware of > with the rest of SURBL. > What's the change? Thanks much for the feedback Daryl and Joseph! I think we probably can't reveal the exact listing criteria in case they're useful for the bad guys. I know it's somewhat inappropriate to ask for comments without revealing details. I suppose I'm asking for general responses then. :) Jeff C.

2 1

RFC: significantly reducing FPs on OB
by Jeff Chan 12 Jul '08

12 Jul '08

We can probably significantly reduce the false positives on ob.surbl.org, the SURBL list based on Outblaze's URI blacklist: http://www.surbl.org/lists.html#ob at the cost of some possibly minor false negatives. SpamAssassin has live (weekly?) statistics about the performance of their rules, including all SURBL lists, against their ham and spam corpora at their Rule QA site: http://ruleqa.spamassassin.org/ As you can see OB ranks significantly below the other SURBL lists, with much higher FP rates around 0.1% compared to 0.01% to 0.025% or so. (Note that the Rule QA site seems to have occasional glitches, so if the numbers seem out of range one week, check again later.) Should we make that change? Jeff C.

3 2

2025

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

2005

2004

Discuss July 2008