Brute force and scanning signatures

[u2521]

Box: PA-2020 (probably all)

OS Version: PANOS-3.0.6

Hi!

I have some questions and suggestions for the threat-ips part of the box.

1. Ability to view settings for scanning and brute force signatures.

- What are the settings for these signatures? I cant find for instance how many attempts/ips/ports is needed for the signature to trigger.

2. Ability to change settings on brute force and scanning signatures.

- Sometimes you want to change the default behaviour of a signature. For instance we want to change how many attempts a client needs to do before a brute force signature triggers. This is also applicable for scanning signatures.

3. More detail in signature summaries

- When looking at a SSH Brute force alarm, it states "count": 1. If its only 1 attempt its not a brute force attack When checking the packet-capture data we only see 1 packet. It would be great to get more information of these "summarizing, threshold" signatures.

4. Regexp signatures

- The ability to view the regular expressions on signatures that you don't need to protect would be nice.

//Henrik

[fredallee · Accepted solution]

Regarding the brute force signatures, we are working on opening them

up the signatures to allow administrators to change how many attempts

a client needs to do over a customizable period of time to trigger a

brute force signature. Right now, we don't expose those thresholds in

the product. So here they are:

1. 40001 :

If a session has the same source and destination and triggers a single

login/authentication event 10 times in 60 seconds, we will identify it

as a brute force attack.

2. 40003 :

If a session has the same source and destination and triggers a single

login/authentication event 100 times in 60 seconds, we will identify

it as a brute force attack.

3. 40004 :

If a session has the same source and destination and triggers a single

login/authentication event 14 times in 60 seconds, we will identify it

as a brute force attack.

4. 40005 :

If a session has the same source and destination and triggers a single

login/authentication event 20 times in 60 seconds, we will identify it

as a brute force attack.

5. 40006 :

If a session has the same source and destination and triggers a single

login/authentication event 100 times in 60 seconds, we will identify

it as a brute force attack.

6. 40007 :

If a session has the same source and destination and triggers a single

login/authentication event 10 times in 60 seconds, we will identify it

as a brute force attack.

7. 40008 :

If a session has the same source and destination and triggers a single

login/authentication event 25 times in 60 seconds, we will identify it

as a brute force attack.

8. 40009 :

If a session has the same source and destination and triggers a single

login/authentication event 10 times in 60 seconds, we will identify it

as a brute force attack.

9. 40010 :

If a session has the same source and destination and triggers a single

login/authentication event 20 times in 60 seconds, we will identify it

as a brute force attack.

10. 40011 :

If a session has the same source and destination and triggers a single

login/authentication event 10 times in 60 seconds, we will identify it

as a brute force attack.

11. 40012 :

If a session has the same source and destination and triggers a single

login/authentication event 7 times in 60 seconds, we will identify it

as a brute force attack.

12. 40013 :

If a session has the same source and destination and triggers a single

login/authentication event 10 times in 60 seconds, we will identify it

as a brute force attack.

13. 40014 :

If a session has the same source and destination and triggers a single

login/authentication event 20 times in 60 seconds, we will identify it

as a brute force attack.

14. 40015 :

If a session has the same source and destination and triggers a single

login/authentication event 20 times in 60 seconds, we will identify it

as a brute force attack.

15. 40016 :

If a session has the same source and destination and triggers a single

login/authentication event 10 times in 20 seconds, we will identify it

as a brute force attack.

16. 40017 :

If a session has the same source and destination and triggers a single

login/authentication event 10 times in 60 seconds, we will identify it

as a brute force attack.

17. 40018 :

If a session has the same source and destination and triggers a single

login/authentication event 40 times in 60 seconds, we will identify it

as a brute force attack.

18. 40019 :

If a session has the same source and destination and triggers a single

login/authentication event 10 times in 20 seconds, we will identify it

as a brute force attack.

Also, I'm not sure I understand question #4 below. If you could

explain a little bit more of what you're looking for, that would be

great.

Thanks,

Alfred

[axel.bodart]

Thank you soooo much, I've been looking for these for days !

While we are at it, how does denying such signature will work ? I guess the deny will be triggered after one minute (when the signature will) and starts dropping further attemps. But if the attacks stops and starts again after a few sec (or even just one), will the first attemps be allowed again for a minute ? Or am I totally wrong ?

[fredallee]

Once the brute force signature threshold is reached will drop all

further attempts for sampling time frame. So if a signature monitors

over 60 seconds, we will drop all further attempts for 60 seconds.

After that, we will start allowing until the threshold is reached again.

Alfred

[u2521]

Hi

Thank you for the reply. Im glad you are working on getting the settings into the product

Some clarification for #4

I'm talking about the signatures that are in the product that look for data patterns for example some sort of buffer overflow. In other products like Cisco, Juniper etc etc you are able to view the "matching pattern" (in regular expression format) for some signatures. Some of the signatures are "protected" because the vendors have agreements with Microsoft etc etc so the vendors are not allowed to give the matching patterns out.

What i'm asking for is the ability to view "what a signatures matches" for all signatures except those you need to protect (encrypt). Also the possibility to change the "matching patterns" on a signatures provided by you would be totally awesome!!

The basic use for that function would be to tune the signatures ourself until we have had time to contact you about a signature that generates alot of false positives and you have tuned the signatures and sent out a new update

//Henrik

[fredallee]

I understand. We have a closed system where we don't show any of the

signatures on our box except for the custom created regex signatures

that are created by administrators. The only method that we have for

fixing signatures is currently through us. So if there are false

positives, please let support know and we will tune the signature and

release a new update. In the meantime, customers can create an

exception for that signature so that it's effectively disabled.

Alfred

[axel.bodart]

Thanks for the explanation. I have another related question : currently we are blocking threats from the level high and above as the medium level seems to be a bit to harsh a level to drop packets at. Since all thoses bruteforce attemps are medium, they are allowed. Is there a way to eitherraise the bruteforce vulnerabilities level to high or add them as blocked ?

Thansk alot !

[fredallee]

You could do this via a custom vulnerability profile, but the drawback

is that you need to manually add newly released signatures.

Alfred

[axel.bodart]

Thansk alot, that would indeed be a bit awfull to manage. So I'm left with 2 options :

1. Customize the severity level for a vulnerability (like applications), bit that's not possible, is it ?

2. Block all medium severity vulnerabilites. Would'nt that lead to too many blocked apps ?

Thaks for advises.

[fredallee]

1. Customized severity for threats is not currently possible.

2. Blocking all medium severity vulnerabilities shouldn't block apps.

If you want to block the brute force attacks, then I would turn on

blocking for medium severity for server side vulnerabilities.

Alfred

[klomboy]

For me it is a must to be able to at least be able to customize the severity level for a non-custom vulnerability signature. I understand that PA might not want to allow modification of their internal signatures, but modifying the severity is a very important requirement for my organization. I bet it is the same for many other organizations. After all, what it is medium for some it could be high for others, or the other way around. Everybody has a different risk tolerance.

By the other hand, it would be very useful to be able to have the ability of making a copy of an internal signature and be able to modify it to suit the customer's needs. In that way organizations can leverage the work already made by PA.

I strongly request to add the ability to customize the severity level as a feature in next releases of PAN-OS.

Thanks.

[u2521]

Any news regarding the ability to view the "matching pattern" or settings for brute force and scanning signatures directly in the GUI ? Will this be available in future versions?

The ability to view the "matching pattern" is very helpful in deciding if the alarm is a false positive or not. If you have pcap active and can see the "matching pattern" you can make up your own mind if its a FP or not.

//Henrik

[panse]

I just ran into the same requirement ... OfficeMax would like the ability to customize the severity level of a threat in our Threat DB.  We already allow them to customize the Risk Level of an App ... they want the same ability for the Severity Level of a Threat/Spyware.

Dave Klein

[swhyte]

Hello Panse,

you can get a hold of your sales contact or call into support and open up an ehancement request regarding this.

thanks,

Stephen