Re: [Pce] draft-ietf-pce-stateful-pce : clarifying the End Of Synchronization marker

[Robert Varga <[e-mail]>]

On 06/21/2016 05:18 PM, [e-mail] wrote:
> Hi,
> 
> Doing some interop testing between two vendors we falled into misinterpretation of the current text of the End Of Sync marker content.
> 
> Here is the current text :
> 
> "The end of synchronization marker is a PCRpt message with the SYNC
>    Flag set to 0 for an LSP Object with PLSP-ID equal to the reserved
>    value 0 (see Section 7.3).  The LSP Object does not include the
>    SYMBOLIC-PATH-NAME TLV in this case, it will include an empty ERO as
>    its intended path and will not include the optional RRO object in the
>    path.  If the PCC has no state to synchronize, it will only send the
>    end of synchronization marker."
> 
> The current text, IMO, has the following issues :
> - it uses non normative wording : "does not include", "will include" , "will not include". How do we need to interpret it ? MUST, SHOULD, MAY ?
> - it does not precise if it can include or not some other objects : can it include an LSP-Identifier object (with all fields to 0) ?

The intent here is to use a minimal PCRpt message, hence we explicitly
exclude SYMBOLIC-PATH-NAME TLV and RRO. ERO is kept empty for the same case.

I think we have not precluded other TLVs from appearing in EOS to allow
future extensions.

I do not think LSP-IDENTIFIERS TLV should be carried here, as it serves
no purpose and is not required -- section 7.3.1's MUST condition does
not trigger, as PLSP-ID=0 is a reserved value and does not identify an LSP.

> It would be good to enhance the text to better describe the content of EOS.
> 
> We suppose that in case there is an issue with the encoding of the EOS marker, the following behavior will be applied, could you confirm ? (typically bad encoding of EOS marker) :
> " The PCE does not send positive acknowledgements for properly received
>    synchronization messages.  It MUST respond with a PCErr message with
>    error-type 20 (LSP State Synchronization Error) and error-value 1
>    (indicating an error in processing the PCRpt) (see Section 8.5) if it
>    encounters a problem with the LSP State Report it received from the
>    PCC and it MUST terminate the session."

Yes. This would trigger, for example, for PLSP-ID=0 and non-empty ERO.

Bye,
Robert