Hi Roman,

Thanks for the review, please see inline

On Thu, 1 Aug 2024 at 20:26, Roman Danyliw via Datatracker <[email protected]>
wrote:

> Roman Danyliw has entered the following ballot position for
> draft-ietf-opsawg-mud-tls-15: No Objection
>
> When responding, please keep the subject line intact and reply to all
> email addresses included in the To and CC lines. (Feel free to cut this
> introductory paragraph, however.)
>
>
> Please refer to
> https://www.ietf.org/about/groups/iesg/statements/handling-ballot-positions/
> for more information about how to handle DISCUSS and COMMENT positions.
>
>
> The document, along with other ballot positions, can be found here:
> https://datatracker.ietf.org/doc/draft-ietf-opsawg-mud-tls/
>
>
>
> ----------------------------------------------------------------------
> COMMENT:
> ----------------------------------------------------------------------
>
> Thank you to Christer Holmberg for the GENART review.
>
> ** Section 1.
>    [malware] indicates that there are observable differences in
>    how malware uses encryption compared with how non-malware uses
>    encryption.  There are several interesting findings specific to
>    (D)TLS which were found common to malware:
> ...
> [bulleted lists of findings]
>
> The techniques, tactics, and procedures (TTPSs) described in the [malware]
> are
> from 2016, 8 years ago.  Is the WG confident that these exact TTPs are
> still in
> use as described?  Will this text age well and is it needed?  The general
> observation that (D)TLS protocols parameters can be used to block malicious
> traffic or traffic that violates policy still stands without the specifics
> having to be described here.
>

Agreed, we will generalize the description of TTPs and focus on the broader
observation that (D)TLS protocol parameters can be leveraged to block
malicious and security policy-violating traffic.



>
> ** Section 1
>    If observable (D)TLS profile parameters are used, the following
>    functions are possible which have a positive impact on the local
>    network security:
>
> What is a “profile parameter”?  How does a “profile parameter” differ from
> any
> other (D)TLS protocol parameter?  Editorially, the idea of profiles hasn't
> been
> introduced yet.
>

Yes, I removed the word "profile" to avoid confusion.


>
> ** Section 1 and 3.  The motivation for this work is repeated multiple
> times as
> being malware and malicious behavior.  Why is more basic enforcement of
> policy
> not also part of this justification.  This is hinted at by text in Section
> 1
> about “ensur[ing] TLS certificates” are valid.  Observing the network for
> use
> of legacy ciphersuite or non-conformant certificates are also common cyber
> hygiene use cases.
>

Good point, updated sections 1 and 3 to high-light the above point.


>
> ** Section 3.
>   Typically, IoT devices have an
>    infrastructure that supports a rapid deployment of updates, and
>    malware agents will have a near-impossible task of similarly
>    deploying updates and continuing to mimic the TLS behavior of the IoT
>    device it has infected.
>
> Realizing that “IoT” can mean number of things, I am surprised by this
> assertion.  I was under the impression that update practices vary for IoT
> devices and rapid deployment could NOT be assumed.  Specifically, being
> deployed on disadvantaged networks or having inconsistent/poor long-term
> support from their manufacturer were among a few of the reasons why timely
> patching of IoT doesn’t happen.
>

Agreed, but this specification is targeted at manufacturers who support MUD
and they would have security as a key criteria. They are more likely to
implement frequent updates to address vulnerabilities.


>
> ** Section 4.1
>    To obtain more visibility into negotiated TLS 1.3 parameters, a
>    middlebox can act as a (D)TLS 1.3 proxy.  A middlebox can act as a
>    (D)TLS proxy for the IoT devices owned and managed by the IT team in
>    the Enterprise network and the (D)TLS proxy must meet the security
>    and privacy requirements of the organization.
>
> If I’m understanding correctly, the setup described above is a full
> terminating
> proxy. At that point, one could certainly look at parameters, but one could
> also perform full content inspection.  It might be useful to mention
> that.  I
> am not sure what this means relative to MUD.
>

The impact of content inspection is discussed in the second paragraph of
Section 9 (Privacy considerations) of the draft.


>
> ** Section 4.2.  What is the (D)TLS profile information being conveyed in
> this
> section?
>

This section refers to enforcing MUD policies at both the DNS (RFC8520) and
TLS levels to enhance overall security; added text for clarity.



> ** Section 5.
>    The (D)TLS profile YANG module provides a method for
>    network security services to observe the (D)TLS profile parameters in
>    the (D)TLS handshake to permit intended use and to block malicious
>    behavior.
>
> I’m having difficulty with the clarity of this language.  How does this
> module
> provide a method to observe anything?  There are no RPC here. Would it be
> more
> accurate to say:
>
> NEW
> This YANG module provides a means to characterize the (D)TLS traffic
> profile of
> a device.  Network security services can use these profiles to permit
>
conformant traffic or to deny traffic from devices that deviates from it.
>

Proposed text looks good, updated draft.


>
> ** Section 8
>    Although it is challenging for a malware to mimic the TLS behavior of
>    various IoT device types and IoT device models from several
>    manufacturers, malicious agents have a very low probability of using
>    the same (D)TLS profile parameters as legitimate agents on the IoT
>    device to evade detection.
>
> Could the observation being made here be clarified?  Is it saying that it
> is
> challenging to mimic TLS behavior across devices, but it could still be
> done by
> malware?  If so, it would be beneficial to describe why this is generically
> true.  Certainly, some of the profile parameters are harder to mimic (e.g.,
> certificate-authorities), but others would not (e.g.,
> signature-algorithm.  The
> difficulty of replications seems tied in some part to how the profile is
> specified.
>

Good point, updated draft.


>
> ** Section 9
>    The MUD URL can be stored in Trusted Execution
>    Environment (TEE) for secure operation, enhanced data security, and
>    prevent exposure to unauthorized software.
>
> It is common for “IoT” to have a TEE?
>

It is becoming common for IoT devices to have TEE, For instance, LAKE WG is
discussing attestation of EDHOC peers (and it is part of the charter) .


>
> ** Section 9
>    However, the
>    malware on the IoT device won't be able to establish a (D)TLS
>    connection with the C&C server to reveal this information because the
>    connection would be blocked by the network security service
>    supporting this specification.
>
> Perhaps soften this language because while it is desired/intended for the
> malware to be blocked the efficacy of this signature based approached is
> not
> guaranteed.
>

Yes, fixed.

Cheers,
-Tiru
_______________________________________________
OPSAWG mailing list -- [email protected]
To unsubscribe send an email to [email protected]

Reply via email to