I hope the logs are first compressed and then encrypted. Otherwise, if
encrypted first then no compression would be achieved.


On Nov 15, 2007 9:24 PM, Michael Starks <[EMAIL PROTECTED]> wrote:
>
> Hello Branko,
>
> > 1. How are logs sent to the server (complete logs or diffs)?
>
> Every event is sent to the server.  OSSEC then evaluates the message
> against the rules.  If there is a match, it is logged and action is
> taken based on the logic in the rule.  If there is no match, then the
> message is discarded (by default, but this can be changed).
>
> Are all the
> > logs that ossec agent is configured to monitor processed by server or
> > agent has some kind of intelligence? I'm asking this because I would
> > need the information about possible network and server load in case when
> > agents are installed servers with big traffic. For example, web servers
> > with big number of visits (which for sure generate big access and
> > probably error logs).
>
> As stated, every event is sent to the server for analysis.  But the
> events are also encrypted and compressed, so the traffic isn't as much
> as the raw event would be.
>
> > 2. Were are the referent integrity checking hashes kept and where does
> > the hash comparison take place?
>
> To the best of my understanding, there is a local database for cache
> purposes (at least there is on Windows), but again, the event is sent to
> the server for comparison with stored checksums.
>
> > 3. How exactly does the rootkit detection work?
>
> There are several things going on here.  One novel method is that OSSEC
> looks at what the OS reports as listening on the host, then attempts to
> bind to every port.  If they don't match up, an alert is sent.  There
> can be problems with this on some installation which open and close
> ports very quickly.  OSSEC uses other methods, and for those I will
> simply refer you to the mailing list archive for a thorough discussion.
>
> HTH,
> Mike
>

Reply via email to