nanog mailing list archives

Re: sub $500-750 CPE firewall for voip-centric application


From: Aris Lambrianidis <effulgence () gmail com>
Date: Fri, 06 May 2016 20:50:36 +0200

Mel Beckman wrote:
The question of code quality is always a difficult one, since in FOSS it’s public and often found lacking, but in private source you may never know. In these cases I rely on the vendor’s public statements about their development processes and certifications (e.g., ICSA). Commercial products often disclose their development processes and even run in-house security threat research groups that publish to the community.

There are also outside certifications. For example, www.icsalabs.com <http://www.icsalabs.com> lists certifications by vendor for those that have passed their test regimen, and both Dell SonicWall and Fortinet Fortigate are shown to be current. PFSense isn’t listed, and although it is theoretically vetted by many users, there is no guarantee of recency or thoroughness of the test regimen.

This brings up the question of whether PFSense can meet regulatory requirements such as PCI, HIPAA, GLBA and SOX. While these regulatory organizations don’t require specific overall firewall certifications, they do require various specific standards, such as encryption strength, logging, VPN timeouts, etc. I don’t know if PFsense meets these requirements, as they don’t say so on their site. Companies like Dell publish white papers on their compliance with each regulatory organization.
It seems those certifications are not offering the assurance at least *some* people would expect from them, unless of course we're talking about feeding the paper pushing beast. This is a mere observation on my part, principally I'm not against them, but I seriously doubt bad coding practices happen only on non certified/audited code, so
I find the question of value difficult to answer in a satisfactory manner.

Random germane example: http://opensslrampage.org/post/83555615721/the-future-or-lack-thereof-of-libressls-fips

Aris


Current thread: