Protocol benchmarking / auditing inquiry

[Brent Kimberley]

Hi.
Can anyone point me to some methods to benchmark and/or audit Kerberos v5?

For example, SSH:
Manual
Read the RFCs and specs.
Semi-automatic.
jtesta/ssh-audit: SSH server & client security auditing (banner, key exchange, encryption, mac, compression, compatibility, security, etc) (github.com)<https://github.com/jtesta/ssh-audit/>
Automatic
SSH Configuration Auditor (ssh-audit.com)<https://www.ssh-audit.com/>


TLS example upon request.

THIS MESSAGE IS FOR THE USE OF THE INTENDED RECIPIENT(S) ONLY AND MAY CONTAIN INFORMATION THAT IS PRIVILEGED, PROPRIETARY, CONFIDENTIAL, AND/OR EXEMPT FROM DISCLOSURE UNDER ANY RELEVANT PRIVACY LEGISLATION. No rights to any privilege have been waived. If you are not the intended recipient, you are hereby notified that any review, re-transmission, dissemination, distribution, copying, conversion to hard copy, taking of action in reliance on or other use of this communication is strictly prohibited. If you are not the intended recipient and have received this message in error, please notify me by return e-mail and delete or destroy all copies of this message.

[Brent Kimberley]

Preferably something smaller and more focused than nmap or OpenSCAP. 😉

[Christopher D. Clausen]

I have used this as a guide, but I think MIT Kerberos version 1.10 is
the latest available:
https://www.cisecurity.org/benchmark/mit_kerberos

Not sure if this is what you are looking for or not.

<<CDC

[Brent Kimberley]

Hi Christopher.

Yes. You are correct. Peer reviewed installation readiness documents like the CIS MIT benchmark are a good "first step."

I was asking pointers to the rest of the lifecycle suite - specifically "walk".

Crawl
=====
Installation readiness documents
e.g., CIS MIT Kerberos Benchmark

Walk
====
Focused applications.

Application which can connect to a client or a server and emit:
Enabled ciphers.
Enabled MACs.
Enabled Kerberos modes (krb5, krb5i, krb5p)
etc.

Background: most sites appear to be misconfigured.

Run
====
A focused service.


-----Original Message-----
From: Christopher D. Clausen <ccla...@acm.org>
Sent: Wednesday, February 14, 2024 2:10 PM
To: Brent Kimberley <Brent.K...@Durham.ca>; kerb...@mit.edu
Subject: Re: Protocol benchmarking / auditing inquiry

[You don't often get email from ccla...@acm.org. Learn why this is important at https://aka.ms/LearnAboutSenderIdentification ]

> (ssh-audit.com)<http://ht/
> tps%3A%2F%2Fwww.ssh-audit.com%2F&data=05%7C02%7CBrent.Kimberley%40Durh
> am.ca%7C8eddde16708448e6cdb008dc2d907d49%7C52d7c9c2d54941b69b1f9da198d
> c3f16%7C0%7C0%7C638435345797172606%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4
> wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C0%7C%7C%7C&s
> data=ydwY2y5%2FxuZxJavbNQw877yOmuFuVo3DktJr%2FdFA05A%3D&reserved=0>

[Brent Kimberley]

To the best of my knowledge" Krb5i provides integrity whereas Krb5p provides confidentiality, integrity, and replay protection.

"Walk tool" finding could map to a radar chart.

In other news, Matthew Palko plans to modernize authentication.
https://techcommunity.microsoft.com/t5/windows-it-pro-blog/the-evolution-of-windows-authentication/ba-p/3926848

[Brent Kimberley]

Minor comment the CIS Benchmark appears to have been written from the system administrator's frame of reference - not the network frame of reference (FoR).
Typically, each frame of reference (FoR) needs to be audited. Hence the need for automation.

[Ken Hornstein]

>Minor comment the CIS Benchmark appears to have been written from the
>system administrator's frame of reference - not the network frame of
>reference (FoR). Typically, each frame of reference (FoR) needs to be
>audited. Hence the need for automation.

I can only say this:

- I've been doing Kerberos for a few decades (but I'm certainly not the
person with the most Kerberos experience on this list).
- I've done a ton of security accreditation work at my $DAYJOB, which
also involves Kerberos. As part of the accrediation work we (and
others) do automated scanning that includes the Kerberos servers
and this seems to satisfy the powers that be. Some of the scanning
seems to detect Kerberos but I am unclear how much it actually checks
for other than "Kerberos is found".
- I've used the aforementioned CIS Benchmark.
- I really have no clue what you mean by "frame of reference" in this
context, and this corresponds to no security accreditation or auditing
requirements I have ever encountered so I cannot provide any
suggestions; I'm really unclear what you are asking for.

--Ken

[Brent Kimberley]

Ken.
The term Frame of Reference is a Cyber Physical system (CPS) term.

For those who work in the cyber subset, the term is "interface".

Regardless of what you call it.

You take the system diagram and evaluate using each major interface or Frame of Reference.

The STIG or CIS benchmark is just one of the interfaces evaluated.


-------------

-----Original Message-----
From: Brent Kimberley

Sent: Wednesday, February 14, 2024 3:24 PM
To: Christopher D. Clausen <ccla...@acm.org>; kerb...@mit.edu

Subject: RE: Protocol benchmarking / auditing inquiry

Minor comment the CIS Benchmark appears to have been written from the system administrator's frame of reference - not the network frame of reference (FoR).
Typically, each frame of reference (FoR) needs to be audited. Hence the need for automation.

[Brent Kimberley]

This approach is taught in first year engineering.

[Brent Kimberley]

At higher levels it falls under "Non Destructive testing".

[Ken Hornstein]

>This approach is taught in first year engineering.

Geez dude, no need to drag me; I'll be the first one to admit that I'm old
and don't know everything! Back in my day our curriculums didn't cover
any computer security topics at all.

But I stand by my original statements: I, personally, have not encountered
those terms before and I've feel it's fair to say I've done a large amount
of accreditation and audit work and some of it involves Kerberos. And
even with your followup emails I'm still unclear what you are asking for.
Is this because I am old and don't know everything? Certainly! Maybe
it's like Zero Trust Security and I am already mostly doing it. Maybe
it's something we have never been asked to do, so I've never done it
(because in the accreditation world you don't seem to get extra credit
for doing stuff that the accreditors do not ask for).

--Ken

[Brent Kimberley]

The purpose of non-destructive testing is to validate form/fit/function - across the entire operational mission/ asset lifecycle/ whatever - contrasted with the STIG/CIS benchmark which throws the real problems "over the wall" to Ken H.

Using the outputs, the lifecycle manager constructs their budget for operations + maintenance (OpEx) and replacement (CapEx).
Physical systems wear out. (Weibull)
Cyber systems fail spectacularly.
CPS systems wear out + fail spectacularly. (Power-law?)

Why is this relevant?

Back in the 1940s, too many planes were falling out of the sky. (Q. How many planes are too many?)
You call this philosophy a "surety system", "fly fix fly", "patch Tuesday", " FAA's approach to the Boeing 737 MAX" - whatever.
Regardless, by the 1950s, it was decided that action needed to be taken. The status quo was unacceptable. It was too expensive for operators.

The national safety council created something called the "Hierarchy of Controls." It was immensely successful. (Planes stopped falling out of the skies.)

You can call this approach "safety by design". This approach and it's benefits are very well documented and might even be applicable to Navy C4ISR.

To tie a bow on this thread:
How can we make Kerberos safe?

-----Original Message-----
From: Brent Kimberley
Sent: Thursday, February 15, 2024 12:19 PM
To: kerb...@mit.edu; ke...@cmf.nrl.navy.mil
Subject: RE: Protocol benchmarking / auditing inquiry

At higher levels it falls under "Non Destructive testing".

-----Original Message-----
From: Brent Kimberley
Sent: Thursday, February 15, 2024 12:12 PM
To: 'kerb...@mit.edu' <kerb...@mit.edu>; 'ke...@cmf.nrl.navy.mil' <ke...@cmf.nrl.navy.mil>
Subject: RE: Protocol benchmarking / auditing inquiry

This approach is taught in first year engineering.

[Brent Kimberley]

Correction:
- Physical systems tend to wear out + fail spectacularly.
- Cyber systems tend to fail silently + inconveniently
- CPS systems tend to wear out + fail spectacularly + fail silently + inconveniently (case in point colonial pipeline)

The purpose of said tools is to evaluate & maintain asset health - over time. (PDCA)