Edwin,

Your solutions would show me a) the flows and b) the octets for 
destination addresses. I was more interested in seeing how *many* 
of this group of internal hosts had connected to the same external hosts.

With the curret flow-tools, I would have to post-process the flows to 
obtain this I think.

C

On Mon, 13 Sep 2004, Edwin Lok wrote:

> Hi Craig,
> 
> You could use flow-nfilter.
> 
> create a file filter.cfg
> 
> # start file filter.cfg
> filter-primitive infected-hosts
>       type ip-address
>       permit <ipaddr>
>       .
>       .
>       permit <ipaddr>
>       default deny
> 
> filter-primitive external-hosts
>       type ip-address
>       permit <ipaddr>
>       default deny
> 
> filter-definition find-infected-hosts
>       match src-ip-addr infected-hosts
>       match dst-ip-addr external-hosts
> # end file filter.cfg
> 
> then:
> flow-nfilter -f filter.cfg -F find-infected-hosts < flowfile | flow-print -f5
> 
> or if you want to get the stats of the src and dst address sorted by octets
> flow-nfilter -f filter.cfg -F find-infected-hosts < flowfile | flow-stat -f10 -S3
> 
> Rgds
> Edwin
> 
> On Sun, 12 Sep 2004, Craig Macdonald wrote:
> 
> > Hi flow-tools people!
> >
> > Background:
> > We've been investigating a virus lookup on campus last week, in light of
> > recent news from the Internet Storm Center.
> >
> > We detected the virus on Friday using our flowscan modules that were
> > processing the flow-tools files.
> >
> > Last week we noted that the virus attempted to connect to 203.81.40.172 on
> > port 10009 (as per http://isc.sans.org/diary.php?date=2004-09-08)
> > So we checked our logs to see which machines had been attempting to
> > connect on that port - this identified most of the infected machines we
> > had previuosly identified and a few more.
> >
> > Question:
> > We were wondering if it is possible, that given a list of (internal) IP
> > addresses, we could filter on them, and then aggregrate to determine if
> > there was anything else that many of the listed machines had attempted to
> > connect to. (Using flow-report etc)
> >
> > If I were to explain this in a SQL like fashion:
> > SELECT dstip, count(*) FROM flows
> > WHERE
> > srcip IS IN (infected hosts list)
> > AND dstip IS external
> > GROUP BY dstip
> >
> > We could then monitor for other infected machines making connections to
> > the same machine (bot controller), thus identifying infections quickly.
> >
> > All input welcome
> >
> > Regards
> >
> > Craig Macdonald
> > [EMAIL PROTECTED]
> > Glasgow University Computing Service
> >
> > _______________________________________________
> > Flow-tools mailing list
> > [EMAIL PROTECTED]
> > http://mailman.splintered.net/mailman/listinfo/flow-tools
> >
> 

_______________________________________________
Flow-tools mailing list
[EMAIL PROTECTED]
http://mailman.splintered.net/mailman/listinfo/flow-tools

Reply via email to