PA doesn't cover DROWN Attack?

[santonic]

A customer has been warned about DROWN attack (https://drownattack.com/) on one of its servers. As a server is behind PA I thought there was no risk. But searching through signature database I didn't find anything about DROWN attack. I've also checked all CVEs connected with attack (CVE-2015-3197, CVE-2016-0703, CVE-2016-0800) and PA doesn't have signature for any of them! 

Anyone knows about if PA covers this attack? Anyone contacted PA about this already?

[kiwi]

Hi santonic,

This has been reported to support yes.

I'd recommend that you open a case with support to get information on the coverage.

Cheers,

-Kim.

[bmorris1]

Hi Santonic,

My understanding is that this attack requires SSLv2 to be supported? (correct me If I am wrong) I think the worse security risk is to have a server that supports SSLv2, I know the PAN  device does not support this, it doesn't even support v3. I know SSLv2 was removed from OpenSSL toolkit on the 15th feb this year with v1.1.0.

Ben

[santonic]

Yes, I agree. But one of the IPS functions is also to protect old or unpatched servers.

Customer was of course advised to upgrade the server.

[Dulle]

From my understanding of this vulnerability, a signature to cover the weakness might be difficult to create.
Since the exploit collects data and then decrypts it, the vulnerability is not based in a protocol anomaly or an easily detectable exploit.
The collection of data based on cracking a 40-bit RSA key, and can be found over time against the realtive limited variations with 40-bit.
Data collection might be by interception between client and server, and this would be undetectable to both.

It might seem better to update the servers use of TLS/SSL, rather than wait for a usefull signature.

[OtakarKlier]

Would the PAN be able to detet this if it was performing reverse proxy ssl decryption?

Just a thought

[Lucky]

Hi,

decryption would not really help, as Dulle explained intercepting communication over long time is sufficient to exploit this. Decryption can't help if someone is somewhere allowing (knowingly or unknowingly) copying of your traffic as it passes along the way.

[Dulle]

And there came Application and Threat Content Release Notes Version 567, and proved me wrong....

[santonic]

Not really Dulle. The signature doesn't detect any exploit. It just detects use of SSLv2.