Re: [P2PSIP] New RELOAD Draft

[Eric Rescorla <ekr at networkresonance.com>]

Thanks for the comments.

At Wed, 23 Jul 2008 21:56:33 -0800,
Michael Chen wrote:
> Here are some notes I have accumulated so far for the July draft:
> 
> A) Section 6.2.1 should add "network byte-order" to the "idiosyncrasies" 
> of the language.

Fair enough, though of course this is how almost everything in IETF
is done.


> B) In 6.2.2, the 'relo_token' 32-bit value should be 0xD2454c4f not 
> 0xC2454c4f.

Thanks.


> C) In 6.2.2, the description for 'var_list_length' should be 
> "...following THREE length fields..."
> 
> D) In 6.2.3.1, the description for 'error_info' says "This MUST be empty 
> (zero length)
>      except as specified below."  However, nothing in "below" explicitly 
> mention being used
>      in 'error_info'.

It's *far* below. I'll fix.


> E) In 6.2.2.2, "ROUTE-LOG-RESPONSE" should be RESPONSE-ROUTE-LOG per 6.2.2.
> 
> F) Section 6.4.2.1.1 must clarify whether the ICE candidate string is 
> with or without
>      the prefix "candidate:" as defined in draft-ietf-mmusic-ice-19 
> section 15.1.

Right. I suspect we'll want replace this with a new encoding....


> G) Suggestion: In Forwarding Header (6.2.2), move the 'overlay' or 
> 'transaction_id' fields
>      next to each other, so that when computing the signature, both 
> fields can be fed to the
>      signing function in a single block instead of two separate ones.

This doesn't add a lot of performance value as compared to the
hashing and digital signing.

-Ekr
_______________________________________________
P2PSIP mailing list
P2PSIP at ietf.org
https://www.ietf.org/mailman/listinfo/p2psip