spamassassin-users mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Kevin A. McGrail" <kmcgr...@apache.org>
Subject Re: CryptoBL [was: Bitcoin rules]
Date Wed, 31 Oct 2018 16:41:21 GMT
+1.  I had the same thought.
--
Kevin A. McGrail
VP Fundraising, Apache Software Foundation
Chair Emeritus Apache SpamAssassin Project
https://www.linkedin.com/in/kmcgrail - 703.798.0171


On Wed, Oct 31, 2018 at 12:21 PM Henrik K <hege@hege.li> wrote:

> On Wed, Oct 31, 2018 at 03:11:51PM +0000, RW wrote:
> > On Wed, 31 Oct 2018 12:03:38 +0100
> > Daniele Duca wrote:
> >
> > > Hello everyone,
> > >
> > > as said some days ago I started a DNSBL based on abused/malign BTC
> > > addresses. This list is queried by an SA plugin that  takes the md5
> > > hash (I know, outdated algorithm, but good enough for this purpose
> > > IMHO)
> >
> > As I pointed out before hashing isn't needed to avoid FPs on case
> > insensitive matches, and it does make things less transparent in
> > debugging.
> >
> > These addresses contain a 160 bit hash of the public key and a 256 bit
> > validity hash. When you convert an alphanumeric string to lower case
> > you only lose about 13% of the entropy, so the probability that two
> > valid and distinct addresses have a case insensitive match is
> > approximately:
> >
> >   1 in 2^360
> >
> > compare that with the probability of the same md5 hash value:
> >
> >    1 in 2^128
> >
> > and the probability that two wallets have the same address:
> >
> >   1 in 2^160
> >
> >
> > With email address lookups the main reason for hashing was privacy,
> > but that obviously doesn't apply here.
>
> No matter, I will implement BTC (and ETH etc), URL and other imaginable
> "hash bl" checks to HashBL.pm with options for raw/md5/sha1 etc.  Everyone
> can run their BLs then how they wish.  ;-)
>
>

Mime
View raw message