Re: [Roll] WGLC for draft-thubert-6man-flow-label-for-rpl-03

[Philip Levis <[e-mail]>]

On Aug 4, 2014, at 1:44 PM, Brian E Carpenter <[e-mail]> wrote:

> It also says that there may be stateful ways of using the flow label.
> It's almost a matter of taste whether the RPL usage is a change to
> the phrase that you quote, or a stateful usage within an RPL domain
> (reverting to stateless use when leaving RPL). I agree that the draft
> could usefully be explicit about this.

I'm a bit confused -- the MUST statement is in Section 2, entitled "IPv6 Flow Label Specification". There doesn't seem to be any text that suggests this restriction is dependent on stateless operation. My interpretation, based on the title and its placement in the document, is that the text in this section applies to any and all uses of the flow label. How am I reading this wrong? Should I be reading the final MUST ("… MUST choose labels that conform to Section 3 of this specification") in Section 4 to implicitly mean that Section 2 doesn't necessarily apply? Or is it optional for stateless as well?

My (admittedly naive) reading is that Sections 3 and 4 related to how source nodes choose a flow label. (S2: "To enable Flow-Label-based classification, source nodes SHOULD assign each unrelated transport connection and application data stream to a new flow.")  This is distinct from what forwarding nodes do to flow labels, which Section 2 covers. 

Phil