Hi Usama,

Thanks for your response, please see my in-line replies with [WW].
The latest version of this draft is attached. Thank you!


Best Regards,
Wei




         Original
         
       
From: Muhammad Usama Sardar <[email protected]&gt;
Date: 2026-04-13 02:10
To: Wei Wang <[email protected]&gt;, Aijun Wang 
<[email protected]&gt;
Cc: tls <[email protected]&gt;, draft-wang-tls-service-affinity 
<[email protected]&gt;, Mohit Sahni 
<[email protected]&gt;
Subject: [TLS] Comments on draft-wang-tls-service-affinity (was Montreal ...)




Hi Wei,
 

Following up with a suggestion on how to proceed:

At a high level, I think it may be helpful to take a step back       and 
approach this from a security-first&nbsp;perspective rather       than a 
design-first&nbsp;perspective. I understand this sounds       like moving back 
to zero, but if you want me to help you, we need       to move more logically 
and it might really help. However, I want       to be clear that the final 
design may&nbsp;come out to be       completely different from what you have.
 

[WW]: Our original intention in proposing this technical solution is to address 
business issues and, on that basis, examine whether any new security problems 
have arisen. It does not focus on security issues, so I think a design-first 
perspective is more suitable.

Of course, this is just a suggestion and as author of draft, you       have the 
choice -- please feel free to take or leave it. If you       would prefer a 
different direction, I kindly ask you to just let       me know so I can plan 
accordingly.

Is it implemented somewhere? If yes, do they have anything public       on how 
it is implemented? That would be really helpful for me.

 
On 09.04.26 11:50, Muhammad Usama       Sardar wrote:
 
              On 09.04.26 08:10, Wei Wang wrote:
 The protocol design in section           4 has been revised based on TLS 1.3.
I am not sure if this one has been fully&nbsp;addressed. As         an example, 
please note that you still have 4xChangeCipherSpec         messages in Figure 1 
while this message seems to have been         removed in TLS 1.3, "except when 
needed for middlebox         compatibility" [0]. There seems to be no 
explanation in your         draft why you need this message. Please explain 
this in the         text.
 
     So based on [2], please remove 4xChangeCipherSpec from the     handshake. 
Before proceeding with the design, I propose to first     focus on the threat 
model and security goals (see below) to help me     understand what you want to 
achieve.
[WW]: Done.

Since we are not very familiar               with security-related 
technologies, would you like to help               us improve the content of 
security considerations? 

Yes, with great pleasure. I am currently stuck in another         high-stake 
issue. When I get some time, I will update my draft         [3] to explain this 
and then come back to you.
 

I've added some details, examples and reference in Sec. 5.2 and       5.3 
[4,5].&nbsp;Please check it out and let me know if it helps you. If       it is 
not helpful, please do not hesitate to let me know what you       find 
difficult, and I will be happy to add more details. You could       start with 
very high-level requirements and ignore the unless       conditions in the 
first stage etc. Like, think about what is it       that you want to protect 
and what does protect mean (integrity,       freshness, etc.).
 

One thing I am missing in the draft is the threat model. I would       like to 
see clear independent subsections (see my proposed       document format in 
draft-usama-tls-fatt-extension) on the threat       model and security goals 
that you want to achieve.

Security goals must be independent of your current design, like       after 
reading the "Introduction" and "Motivation and Design       Rationale" 
sections, I should be able to jump over directly to the       section on threat 
model and security goals. Can you try it based       on [4] and [5]? and then 
we can work together to improve it.&nbsp;

What do you think? Are you open to this approach?
 

[WW]: Thank you for the detailed explanation and the references.
I find drafting the Threat Model quite challenging, but I have attempted to 
list some Informal Security Goals in Section 7.2. Could you please review them 
and help us further refine the content of both Section 7.1 and Section 7.2 
based on this?
     Best wishes, 
-Usama

[0]  
https://www.ietf.org/archive/id/draft-ietf-tls-rfc8446bis-14.html#section-1.3-2.6.1
 

[1] https://www.ietf.org/archive/id/draft-ietf-tls-rfc8446bis-14.html

[2] https://github.com/tlswg/tls13-spec/issues/1413
 

[3]  
https://www.ietf.org/archive/id/draft-usama-tls-fatt-extension-02.html#section-4.3

[4] 
https://www.ietf.org/archive/id/draft-usama-tls-fatt-extension-03.html#section-5.2

[5] 
https://www.ietf.org/archive/id/draft-usama-tls-fatt-extension-03.html#section-5.3

Attachment: draft-wang-tls-service-affinity-02.pdf
Description: Binary data

Attachment: draft-wang-tls-service-affinity-02.xml
Description: Binary data

_______________________________________________
TLS mailing list -- [email protected]
To unsubscribe send an email to [email protected]

Reply via email to