Hi all, 

I have submitted a draft about minimal TLS in the spirit of the work that Tero 
was doing. 
Working on the document, which you can find at 
http://tools.ietf.org/html/draft-tschofenig-lwig-tls-minimal-00.html, I noticed 
the following issues:

Just providing a writeup that illustrates how to implement a shared secret 
based ciphersuite is not sufficient. The reason is that in certain cases shared 
secret based variants do not meet the security requirements. Hence, when we 
look at more than just a single ciphersuite then the question arises where to 
stop looking at the various extensions. This is a scoping question. 

I don't think it makes sense to replicate the text from the original 
specification. The document cannot violate the original specification itself. 
It can only provide design tradeoffs. 

Ideally, to give engineers more guidance one would have to provide some code 
size indications. Of course specific numbers only serve as an indication/hint 
since the code size depends on the specific platform and the degree of 
optimization applied to the code. 

Finally, there is the question whether codesize is the only aspect to look at. 
What about memory requirements, bandwidth consumption, etc.? 

In a nutshell, to offer valuable guidance this could be a longer exercise. (I 
would like to work with a few others on this topic. Does not sound fun to do 
this alone. Maybe there even a chance to produce a lightweight TLS reference 
implementation or so.)

I wonder what others think. 

Ciao
Hannes

_______________________________________________
Lwip mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/lwip

Reply via email to