2018-03-28 03:35 PM
I have a question from an internal customer:
would like to get RSA engaged and find out how to further tune false positives from the spectrum/malware analysis appliance. For instance if we saw a mcafee dat file fire that was a false positive, we would like to tune this filename/hash /various other characteristics from preventing the tool to fire in the future. If we can get this to tune as requested, would like to ingest syslog into the SIEM and look at potential alerts .
The customer already has RSA Security Analytics User Documentation. Is there any other resource he can access to help answer his questions?
2018-03-29 01:00 PM
I would take a look at these items to see where you have the opportunity to filter out files from the malware pipeline.
most likely you will add an app rule to mark that file type/extension or other combination with a value of spectrum.filter into the content metakey to remove that flow from being sent to the MA appliance.
2018-03-29 01:26 PM
thanks eric. how do we tune this within the spectrum appliance? We shouldn't have to go into and modify tags outside of the appliance just to tune this.
For instance we want to tune out / in filenames/hashes/extensions from iocs that fire here
2018-03-29 01:38 PM
Several things to do to filter down the traffic to Malware Appliance:
2018-03-29 01:41 PM
select * where content='spectrum.consume' || content='spectrum.consume11'
that is the filter command from spectrum, you could try to adjust that but its not scaleable to try to update that for specific filters.
Using a feed or app rule to filter in or out items from the MA pipeline is the best way to do it as far as I know.
In the MA > Config > Hash tab you can search for an MD5 in there and add it to trusted (edit the entry and save it). That way you can trust certain hashes that have come in already.
2018-03-29 02:57 PM
Actually, spectrum parser now only tags sessions with appropriate files as “spectrum.analysis” then the Application Rule applies the filtering:
spectrum.consume content=’spectrum.analysis’ && content != ‘spectrum.filter’
As I noted before, adding && direction=’outbound’ will drastically cut down on the number of sessions even sent to MA, cutting out Lateral (internal-internal) and inbound (external-internal) sessions.
so to filter sessions PRIOR to MA even getting them, use either an app rule, feed, or custom parser to tag those sessions with ‘spectrum.filter’ in the “content” metakey.
This is far more effective than filtering them by hash in the appliance itself.
A Long time ago, I helped create filtering for your deployment Joe, if you are still at the same company. I cut it down from 60000 sessions per hour to 600 sessions per hour being consumed, but I fear all that filtering was lost some time in the past.
John E. Snider | Senior Consultant, SME, Professional Services | GCIH, GCFA | RSA
m: +1 (281) 813-7147 |e: john.snider@rsa.com<mailto:john.snider@rsa.com> | www.rsa.com<http://www.rsa.com/> | Time Zone: America/Central Time (CST6DST)
<http://www.rsa.com/>
<https://community.rsa.com/welcome>