Re: [PWE3] 答复: flow label format in draft-ietf-fat-pw-02

[Stewart Bryant <stbryant at cisco.com>]

Lucy

Thanks for the comments.

Hi Stewart,

Thanks. Flow Label Sub-TLV is for signaling purpose. Figure 3 title is much better.

However text in section 4 is bit confusion.

 

A PE that wishes to use a flow label includes in its label mapping

   message a Flow Label Sub-TLV (FL Sub-TLV) with F = 1 (see

   Section 4.1).  A PE that can correctly process a flow label, and is

   willing to receive one, but does not wish to send a flow label,

   includes an FL Sub-TLV with F = 0.

[LY] egress PE processing a flow label is just discarding it. Right?

Yes, the processing operation is a simple discard, although an implementation might choose to check EOS.

 

   If a PE has sent an FL Sub-TLV with F = 1, and has received an FL

   Sub-TLV it MUST include a flow lablel in the label stack.

[LY] should be “…, and has received an FL Sub-TLV with F=1, egress PE MUST process a flow label in the label stack.”

The original statement is correct. A contracted with B to send an FL (sent FL sub-TLV with F=1). B agreed to the contract (sent FL sub-TLV). Now A must honour the contract by including an FL.

 

   If a PE has sent an FL Sub-TLV with F = 1 and does not receive an FL

   Sub-TLV it MUST send a new label mapping using an FL Sub-TLV with F =

   0.

[LY] This is egress PE.

No - ingress. A tried to negotiate using a FL (FL Sub-TLV with F = 1), B said no (did not send an FL sub-TLV), So A now sends an Fl Sub-TLV with F = 0. If B understand the Sub-TLV it knows that A will not include FL in the data packet. If B does not understand the Sub-TLV it is none the wiser and processes continues to process packets as if FL were never invented.

 

 

   A PE that has sent an FL Sub-TLV with F = 0 MUST NOT include a flow

   label in the label stack.

[LY] is this ingress PE or egress PE?

Ingress

 

   If a PE that previously did not received a label binding without a FL

   Sub-TLV receives a new a label mapping with one included, it MAY send

a new label mapping including an FL Sub-TLV with F = 1.

[LY] This is not clear to me.

If A previously received from B no FL sub-TLV, A had to send a sub-TLV with F=0 and not use a FL (see above). If A now receives an FL  sub-TLV from B, it can change its mind and issue a new  FL sub-TLV with F=1. This is needed to cover the case of single sided signaling, where B may use A's old label binding.
However I could omit this sentence since it's obvious that A can re-issue its label when B changes its mind.

Please use egress PE and ingress PE to describe the actions and make sure all the cases covered.

I will take a look at the text, but since the PEs are both ingress and egress I am not sure that helps. However I will look to see if it can be clarified.

- Stewart