You are viewing a plain text version of this content. The canonical link for it is here.
Posted to users@spamassassin.apache.org by Chris Santerre <cs...@MerchantsOverseas.com> on 2004/04/19 21:10:43 UTC
RE: Test Received headers using CIDR addresses rather than regex
ps
> -----Original Message-----
> From: Jeff Chan [mailto:jeffc@surbl.org]
> Sent: Friday, April 16, 2004 6:43 PM
> To: SpamAssassin Users
> Subject: Re: Test Received headers using CIDR addresses rather than
> regexps
>
>
> On Friday, April 16, 2004, 12:49:07 PM, Matt Kettler wrote:
> > DNSBLs are also inherently by far more scaleable than
> static SA rules. Why?
> > because you don't need to store the entire database on your
> local machine.
>
> I agree, but...
>
> > Witness the overhead of SURBL vs BigEvil.
>
> In some fairness BigEvil represents a much larger dataset than
> sc.surbl.org does currently. (sc.surbl.org has about 500 of the
> most reported SpamCop URI domains while BigEvil probably contains
> the equivalent of thousands of domains.)
I'm sure it is around 4000+, with all the crazy regex that is in there. I
hope to get them dumped out when I find the script and the time.
>
> A better comparison if someone wanted to do two performance tests
> might be to compare Bill Stearns' sa-blacklist in SA rule form
> versus essentially his same data turned into a name-RBL in the
> SURBL ws.surbl.org. In particular using an RBL moves the storage
> of the spam domain data out of SA memory and into the local name
> service cache.
Yup. BE still only increases memory by about a 1.4 megs. That ain't too
shabby ;)
--Chris (Mr. Game MisConduct!)