[Snort-users] v2.8.4 incorrect logging to MySQL

Martin Roesch roesch at ...1935...
Tue Apr 14 16:39:26 EDT 2009


Loyal Moses:

Sourcefire is limited on man power as is any organization that has
limited supplies of time and money.

Sourcefire does not use the spo_database plugin AT ALL in our
commercial products because they impart unacceptable performance
overhead into the product.  We use the Unified2 logging infrastructure
and our own log spooling system in the commercial products.  The
spo_database code was developed and supported by external contributors
to the project back in the early 2000's who abandoned it and left it
for us and the community to maintain.

The biggest support item on the mailing list from the open source
users for many years now has been the database plugin because it's
complexity is so high.  To my mind it makes a lot of sense to reduce
the complexity of Snort configuration and improve the worst case
performance of Snort by removing any code that's a constant
maintenance headache and that doesn't relate to the core functionality
of network traffic inspection.  Less lines of code = less bugs = less
security issues = less code audit = faster time to release code.  In
the limited number of hours that we have to write code I'd rather
spend that time improving the detection capabilities and performance
of Snort than maintaining a zoo of output plugins.

If you've noticed, database connectivity isn't in the Snort 3.0
architecture (e.g. SnortSP Beta 3) at all, nor will it be in the
distribution tarball.  If someone wants to maintain an external patch
to the system that's fine but there's no reason to put it in the
distribution code base.  That's MY call and I think it's a good one
for the reasons I listed above.  Using an external data spooler like
Barnyard/Barnyard2/SnortUnified.pm is a better way because it
maximizes the performance of Snort while lowering its runtime
configuration and code complexity.  You have the additional step of
maintaining the spool directories but the complexity is moved out of
Snort which is a win in my opinion.

Jack Pepper:

I believe the latency between detection and spooling to a database
using the unified/Barnyard method is preferable to losing packets and
missing detections altogether using the direct-to-database method.  We
use Unified2 here at Sourcefire for our commercial devices.

Paul Schmehl:

Now that Barnyard2 is available and being actively maintained I would
consider Barnyard to be deprecated.  There is no active development
going on around Barnyard nor will there be unless someone can talk
Andrew into doing some work on it.  Given that the Barnyard2 group is
active and motivated I would favor that route.  Barnyard2 isn't
"officially supported" but I think it's the better route to take
ultimately.

Danny Paul:

I'd be interested to see your data around that, we haven't seen that
sort of thing here but we use a different database schema and spooling
infrastructure.

Marty


On Tue, Apr 14, 2009 at 3:35 PM, Loyal A Moses <loyalmoses at ...3027...> wrote:
>
> Is Sourcefire limited on development skill or man power?
>
> It makes no sense at all to remove one of the most common facilities
> in use by snort users because it is "too complex".
>
> In the end, you'll do what you are going to do regardless of the
> community -- we've seen it before. But don't use "complexity" and
> "bugs" as the excuse.
>
> Sourcefire is a publicly traded company -- Is it smart to be taking
> votes on product development from a mailing list? I wouldn't think so.
>
> Loyal.
>
> On Apr 14, 2009, at 11:52 AM, Jason Brvenik wrote:
>
>> I have an ulterior motive and it is simple.
>>
>> Many of the bugs and issues over time with snort have been in output
>> plugins. Make one well supported, tested, unified method designed for
>> best performance and while doing so it improves the supportability and
>> maintainability of the code base.
>>
>> On Tue, Apr 14, 2009 at 2:39 PM, Loyal A Moses <loyalmoses at ...3027...>
>> wrote:
>>> My vote is to provide as many output options as possible, to help
>>> keep
>>> snort used as a tool.
>>>
>>> The argument of code complexity being a good reason to remove output
>>> facilities is only valid if the code is written poorly and not
>>> modular. This wheel doesn't need re-invented and this conversation is
>>> kind of silly, unless there is ulterior motives for actually wanting
>>> to remove this support.
>>>
>>> Loyal.
>>>
>>> ------------------------------------------------------------------------------
>>> This SF.net email is sponsored by:
>>> High Quality Requirements in a Collaborative Environment.
>>> Download a free trial of Rational Requirements Composer Now!
>>> http://p.sf.net/sfu/www-ibm-com
>>> _______________________________________________
>>> Snort-users mailing list
>>> Snort-users at lists.sourceforge.net
>>> Go to this URL to change user options or unsubscribe:
>>> https://lists.sourceforge.net/lists/listinfo/snort-users
>>> Snort-users list archive:
>>> http://www.geocrawler.com/redir-sf.php3?list=snort-users
>>>
>
>
> ------------------------------------------------------------------------------
> This SF.net email is sponsored by:
> High Quality Requirements in a Collaborative Environment.
> Download a free trial of Rational Requirements Composer Now!
> http://p.sf.net/sfu/www-ibm-com
> _______________________________________________
> Snort-users mailing list
> Snort-users at lists.sourceforge.net
> Go to this URL to change user options or unsubscribe:
> https://lists.sourceforge.net/lists/listinfo/snort-users
> Snort-users list archive:
> http://www.geocrawler.com/redir-sf.php3?list=snort-users
>



-- 
Martin Roesch - Founder/CTO, Sourcefire Inc. - +1-410-290-1616
Sourcefire - Security for the Real World - http://www.sourcefire.com
Snort: Open Source IDP - http://www.snort.org




More information about the Snort-users mailing list