Dear WG, Just to inform you that a new revision of PCEP has been posted, with the following main changes: 1) We fixed a number of inconsistencies between MUST, MAY, ... + other semantic inconsistencies here and there, add new references to the application-specific requirement ID for reference, clarified the text in a number of places, clarified the notion of optional/mandatory objects. 2) The ID has been restructured a little bit: the Elements of Procedure related to several objects were defined in a separate section 9, making the ID not easily readable. 3) Appendix A has been removed since it was essentially useful during the protocol selection process. Some substantial changes ******************************* 1) We had introduced two modes: permanent versus per-request. We actually do not need to specify this and negotiate the mode upon session establishment. If the PCC desires to use a per-request mode and establishes a new PCEP session on a per-request basis, this is by default supported. So we updated the ID accordingly. 3) Close message + CLOSE Object added. 4) We added a Deadtimer field in the Open message rather than fixing the Deadtimer to x times the Keepalive since this might be quite useful in several circumstances. 5) The PCEP Session establishment phase with parameter negotiation has been updated. Now a PCEP peer can propose alternative PCEP session characteristics value by including an OPEN message in PCErr. Note that this is also reflected in the FSM. 6) The DELAY object has been removed. A new object (more generic) and named the METRIC object has been defined that can be used within a PCReq to specify the metric that should be optimized and also a bound (optionally) and in a PCRep to indicate the computed path metric. This introduces a new more flexible logic (C bit of the RP object has been removed). 7) The logic of the P-flag and I-Flag (in common header) has been rewritten. 8) The structure of the OPEN object has been changed (reduced). R-flag removed. 9) NO-PATH object: section has been reworded for the sake of clarity. Ability to suggest a value instead of just reporting the unsatisfied constraint. Example added. 10) We clarified that the IP address in the END-POINTS object does not have to be the source/destination addresses of the TE LSP since this was discussed on the list. 11) SVEC object: added to ability to send synchronized path computations in multiple PCReq messages (new PCEP variable added). 12) The B and N bits of the LSPA object have been removed. 13) PCRep message structure has been updated Of course, comments are more than welcome ! Thanks. JP. Begin forwarded message:
|
_______________________________________________ Pce mailing list [email protected] https://www1.ietf.org/mailman/listinfo/pce
