Re: [lamps] Call for adoption of draft-housley-lamps-3g-nftypes

[tirumal reddy <[e-mail]>]

On Wed, 10 Aug 2022 at 12:02, Tomas Gustavsson <
[e-mail]> wrote:

> I don't think 3GPP networks will make use of certificate transparency
> logs. These are internal telco networks and will not use publicly trusted
> WebPKI CAs for issuing TLS certificates. I don't think publicly trusted CAs
> could even issue these certificates as it may contain other information
> than what's allowed by Baseline Requirements, such as internal
> hostnames/IPs.
>

Thanks. Please update the draft to say the deployment model uses an
internal CA and not a public WebPKI CA.


>
> There are some guards against malicious network functions built into the
> 3GPP specification, by the usage of vendor certificates for authenticating
> the network functions the MNO plans to put into it's network.
>

A pointer to the relevant document will be helpful.

-Tiru


>
> Cheers,
> Tomas
>
> ------------------------------
> *From:* Spasm <[e-mail]> on behalf of tirumal reddy <
> [e-mail]>
> *Sent:* Wednesday, August 10, 2022 7:22 AM
> *To:* Russ Housley <[e-mail]>
> *Cc:* LAMPS <[e-mail]>
> *Subject:* Re: [lamps] Call for adoption of draft-housley-lamps-3g-nftypes
>
> CAUTION: External Sender - Be cautious when clicking links or opening
> attachments. Please email [e-mail] with any questions.
>
> Hi Russ,
>
> Please see inline
>
> On Mon, 8 Aug 2022 at 21:01, Russ Housley <[e-mail]> wrote:
>
> Tiru:
>
> 1. Yes, this is a good topic to expand the Security Considerations.
>
> 2. This seems pretty obvious to me, but I will think about a sentence or
> two for a more complete explanation.
>
>
> Thanks. You may want to also discuss the privacy and security implications
> of using NFType in the certificate extension for RBAC. For example (1) If
> TLS 1.2 is used by network functions, pervasive monitoring is possible for
> an attacker to identify the NFTypes visible in the TLS handshake and can
> potentially target a specific NFType (e.g., subject to DDoS or launch a
> targeted attack). (3) Misuse of NFType to gain additional privileges and
> what are the potential remediation techniques ?
>
>
> Yes, the certificate is plaintext when TLS 1.2 is used, and it it
> encrypted when TLS 1.3 or IKEv2 is used.
>
>
> In TLS 1.3 (without encrypted client hello), SNI will not be encrypted and
> it is possible for an attacker to get the certificate content from
> certificate transparency logs to identify the NFTypes associated with the
> FQDN.
>
>
>
> I'm not sure what you mean about misuse of the NFType.  Are you talking
> about the trusted CA putting the wrong NFType in the certificate?
>
>
> No, trusted CA may not inject a wrong NFType and it can be validated by
> the network function sending the CSR to the CA.
> I meant the NFTypes and FQDN of network functions will be available in the
> certificate transparency logs. It exposes the internal/external network
> functions details to anyone on the Internet. It may also be possible for an
> internal attacker to host a malicious network function and misuse the
> NFType to gain additional privileges.
>
> Cheers,
> -Tiru
>
>
> Russ
>
>