Re: [v6ops] IPv6 Diagnostic Option

Thanks to all for your comments.  But, there may be some confusion as to what we are trying to achieve.  The Proposed Flow Label field, in our opinion, does not meet our requirements. The reasons are as follows:

1.  The Flow Label, as defined, is a unique and (generally speaking) static identifier per flow.  Its purpose is to distinguish one flow from another.  

2.  Our proposed field increments per packet WITHIN a flow.  The intended purpose is to distinguish one packet from another not one flow from another.

3.  The Flow Label is only 20 bits which is not sufficient to hold even the existing 32 bit ID.  We are proposing a 64-bit ID. 

Nalini Elkins
Inside Products, Inc.
(831) 659-8360
www.insidethestack.com

--- On Thu, 7/21/11, Roland Bless <roland.bless@kit.edu> wrote:

> From: Roland Bless <roland.bless@kit.edu>
> Subject: Re: [v6ops] IPv6 Diagnostic Option
> To: v6ops@ietf.org
> Cc: ipv6@ietf.org
> Date: Thursday, July 21, 2011, 10:41 PM
> Hi,
> 
> On 22.07.2011 06:38, nalini.elkins@insidethestack.com
> wrote:
> > Would love to get feedback on an Internet Draft we
> have submitted.
> 
> > The issue we are addressing is that the IP
> Identification field which
> > was available with IPv4 in the main header is
> available with IPv6
> > only in the Fragment Header.
> 
> I don't think that your proposed solution is necessary,
> since
> IPv6 has the Flow Label, that - if set by the source - may
> be
> very well suited for your purpose (tracking flows along a
> path).
> Please see http://datatracker.ietf.org/doc/draft-ietf-6man-flow-3697bis/
> for more details.
> 
> Kind regards,
>  Roland
> --------------------------------------------------------------------
> IETF IPv6 working group mailing list
> ipv6@ietf.org
> Administrative Requests: https://www.ietf.org/mailman/listinfo/ipv6
> --------------------------------------------------------------------
> 