Re: [Pana] Fwd: New Version Notification for draft-yegin-pana-encr-avp-01.txt

[Robert Cragie <[e-mail]>]

Tanaka-san,

Thank you for your comments. I have responded inline, bracketed by 
<RCC></RCC>.

Robert

On 26/03/2012 7:00 AM, Yasuyuki Tanaka wrote:
> Hi,
>
> I've read the draft and I have four comments. Please see them below.
>
> Best,
> Yasuyuki Tanaka
>
> ---------------------------------------------------------------------
>
> (1) Section 3
> I found a typo in the last sentence.
>
> >   The length of PANA_ENCR_KEY depends on the integrity algorithm in
> >   use.
>
> should be
>
>     The length of PANA_ENCR_KEY depends on the *encryption* algorithm in
>     use.
<RCC>Agreed, good catch.</RCC>
>
> ---------------------------------------------------------------------
>
> (2) Section 4
> I've not found text about key size of AES to be used for
> AES_CTR. Probably 128-bit is assumed.
>
> It would be better to clarify the key size to be used for the
> algorithm and to use "AES128_CTR" or something instead of "AES_CTR".
<RCC>Agreed.</RCC>
>
> ---------------------------------------------------------------------
>
> (3) Section 4
> The definition of "q" is different between the draft and NIST
> SP800-38C. In this draft, "q" is defined as "octet length of message
> length field." But in NIST SP800-38C, "q" is defined as "The octet
> length of the binary representation of the octet length of the
> payload."
>
> In addition, I don't realize what 'length of message length field'
> means... Why is "q" 3? The length of "Message Length" field of PANA
> Message Header is 2 octets...
>
> Anyway, I think the following part might cause some confusion.
>
> >      AES-CTR (Counter) encryption algorithm as specified in
> >      [NIST_SP800_38A].  The formatting function and counter generation
> >      function as specified in Appendix A of [NIST_SP800_38C] are used,
> >      with the following parameters:
> >
> >
> >            n, octet length of nonce, is 12.
> >            q, octet length of message length field, is 3.
>
> IMHO, it's better to say just "q is 3" as the definition of "q".
<RCC>Agreed. It was chosen at 3 to be consistent with the use of CCM in 
the proposed TLS cipher suites.</RCC>
>
> ---------------------------------------------------------------------
>
> (4) Section 4
> It would be very helpful to provide an example of the first counter.
>
> When Key-Id is 0x55667788, Session ID is 0xaabbccdd, and Sequence
> Number is 0x11223344, the correct first counter is
> 0x0255667788aabbccdd11223344000001. is it correct?
<RCC>That is correct. I will include the example as given.</RCC>