[Snort-users] Maximum Number Of IPs Per Variable In snort.conf

Matt Kettler mkettler at ...4108...
Mon Aug 1 19:01:01 EDT 2005


O'Sullivan, Mairtin wrote:
> Apologies if this comes through two times. I sent it a few days ago from
> an account which wasn't a member of Snort-Users.
> 
> I was wondering what's the maximum number of IPs you can have in a
> variable in snort.conf?
> 
> In the post below it states that the performance hit would be too great
> to even attempt introducing a large number of IPs. Has that changed
> since 2002?
> http://archives.neohapsis.com/archives/snort/2002-12/0600.html

AFAIK, no, that hasn't changed.

I also don't think you'll see support for it anytime soon either, as I can't
think of an efficient way to implement it. (But there are many people out there
smarter than me, so bear in mind this is just an opinion)

I suppose you might be able to do some really crazy many-list structure, but it
would be a lot of work and suck up memory.

You'd wind up having a deeply nested series of lists pointing to other lists all
cross-referencing down to the same content rule lists.

You'd start with a list of source-ip specifiers
Those entries would each point to a list of source-port specifiers
Those entries would each point to a list of dest-ip specifiers
Those entries would each point to a list of dest-port specifiers
Those would point to a list of content rules.

That would probably also hurt performance in the single-range case, so I don't
think it would be quite so good for the general snort community.


> At present I was to look at putting roughly 300 /32 addresses into a
> single variable.
> 
> They addresses are not consecutive and so can't be supernetted.
> 
> Any thoughts?

My only thought is why.






More information about the Snort-users mailing list