Re: [secdir] draft-ietf-manet-packetbb-sec-08 SecDir review

[Ulrich Herberg <[e-mail]>]

Dear Donald,

thank you very much for your review of the draft. We have addressed your
comment about section 8.1 and 9.1 in a new revision which we just submitted.

Best regards
Ulrich

On Wed, Feb 29, 2012 at 7:53 AM, Donald Eastlake <[e-mail]> wrote:

> My apologies for getting this review in late.
>
> I have reviewed this document as part of the security directorate's
> ongoing effort to review all IETF documents being processed by the
> IESG.  Document editors and WG chairs should treat these comments just
> like any other last call comments.
>
> Overall, I do not have security concerns with this document. See
> comments below.
>
> This document describes "signature" and "time" building blocks for
> constructing messages/packets as described in RFC 5444. There is
> actually noticeable overlap with Section 7.1 of RFC 5444, enough that
> I am inclined to say that this draft should indicate the it Updates
> 5444.
>
> The Security Considerations Section says "...  has the same security
> considerations as [RFC5444]." In turn, RFC 5444 says "This
> specification does not describe a protocol; it describes a packet
> format.  As such, it does not specify any security considerations;
> these are matters for a protocol using this specification." :-) But,
> in fact, the Security Considerations Section of 5444 continues with
> design suggestions for authentication/integrity and confidentiality.
> Arguably, this draft provides a more detailed syntax with some
> processing rules for authentication/integrity with signatures
> extending 5444. But it still defers much to any specific MANET
> protocol making use of RFC 5444, this draft, and probably additional
> "building block" drafts or RFCs.
>
> It appears that the MANET WG is approaching all this through a series of
> overlapping documents each of which is of limited content but provides
> more details. For example, this draft sets up hash function and
> cryptographic function IANA registries but provides only the identity
> function as initial content for these registries. Presumably additional
> documents will request allocations from these registries for other
> functions. There is nothing inherently wrong with this approach and
> trying to produce large monolithic documents can have problems. But a
> swarm of smaller inter-related and partly overlapping documents can be
> confusing.
>
> Section 8.1 and 9.1: These sections provide that when adding a packet
> or message signature TLV, respectively, any pre-existing packet or
> messages signature "MUST" be removed, etc., before signature calculation
> but
> only "SHOULD" be restored afterwards. I would have guessed that
> "SHOULD" would be a "MUST". In any case, it might be good to say when
> you don't need to restore a signature TLV, which I would assume would
> be if that signature TLV is not needed by the recipient.
>
> Thanks,
> Donald
> =============================
>  Donald E. Eastlake 3rd   [phone] (cell)
>  155 Beaver Street, Milford, MA 01757 USA
>  [e-mail]
>