Updated Rules for PCEP Object Ordering

Status of This Memo

This Internet-Draft is submitted in full conformance with the provisions of BCP 78 and BCP 79.¶

Internet-Drafts are working documents of the Internet Engineering Task Force (IETF). Note that other groups may also distribute working documents as Internet-Drafts. The list of current Internet-Drafts is at https://datatracker.ietf.org/drafts/current/.¶

Internet-Drafts are draft documents valid for a maximum of six months and may be updated, replaced, or obsoleted by other documents at any time. It is inappropriate to use Internet-Drafts as reference material or to cite them other than as "work in progress."¶

This Internet-Draft will expire on 10 January 2023.¶

Copyright Notice

Copyright (c) 2022 IETF Trust and the persons identified as the document authors. All rights reserved.¶

This document is subject to BCP 78 and the IETF Trust's Legal Provisions Relating to IETF Documents (https://trustee.ietf.org/license-info) in effect on the date of publication of this document. Please review these documents carefully, as they describe your rights and restrictions with respect to this document. Code Components extracted from this document must include Revised BSD License text as described in Section 4.e of the Trust Legal Provisions and are provided without warranty as described in the Revised BSD License.¶

Table of Contents

1. Introduction

[RFC5440] describes the Path Computation Element (PCE) Communication Protocol (PCEP). PCEP defines the communication between a Path Computation Client (PCC) and a PCE, or between PCEs, enabling computation of Multiprotocol Label Switching (MPLS) for Traffic Engineering Label Switched Path (TE LSP) characteristics.¶

[RFC5440] defines several PCEP messages. For each PCEP message type, rules are defined that specify the set of objects that the message can carry using [RFC5511]. Further, [RFC5440] states that the object ordering is mandatory. This causes confusion when multiple extensions add new objects in the PCEP messages and the respective order of these new objects is not specified (see [EID6627]).¶

This document updates [RFC5440] to relax the strict object ordering requirement.¶

2. Conventions

The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", "SHOULD", "SHOULD NOT", "RECOMMENDED", "NOT RECOMMENDED", "MAY", and "OPTIONAL" in this document are to be interpreted as described in BCP 14 [RFC2119] [RFC8174] when, and only when, they appear in all capitals, as shown here.¶

3. Motivation

The mandatory object ordering requirement in [RFC5440] is shown to result in exponential complexity in terms of what each new PCEP extension needs to cope with in terms of reconciling all of the previously-published RFCs, and all concurrently work in progress in the form of the internet-drafts. This requirement does not lend itself for the extensibility of PCEP.¶

4. Update to RFC 5440

Section 6 of [RFC5440] states:¶

   An implementation MUST form the PCEP
   messages using the object ordering specified in this document.

This text is updated to read as follows:¶

   An implementation SHOULD form the PCEP
   messages using the object ordering specified in this and
   subsequent documents when an ordering can be unambiguously
   determined; an implementation MUST be prepared to receive
   a PCEP message with objects in any order when possible.

This update does not aim to take away the object ordering completely. It is expected that the PCEP speaker will follow the object order as specified unless there are valid reasons to ignore. It is also expected that the receiver is able to unambiguously understand the object meaning irrespective of the order.¶

5. Compatibility Considerations

While one of the main objectives of the changes made by this document is to enable backward compatibility between PCEP extensions, there remains an issue of compatibility between existing implementations of [RFC5440] and implementations that are consistent with this document.¶

It should be noted that common behavior for checking object ordering in received PCEP messages is as described by the updated text presented in Section 4. Thus, many implementations, will still have implemented a consistent and future-proof approach. However, for completeness, it is worth noting how behaviors might interact between implementations.¶

The messages generated by an implementation of this document when received by a legacy implementation with a strict interpretation of object ordering MAY lead to error handling. It is interesting to note that the [RFC5440] does not define an Error-Type and Error-value corresponding to this error condition.¶

6. Open Questions

• Should a new flag or a TLV in Open Message be added to exchange this capability? Not sure if this is strictly needed if we can live with Section 5.¶

7. Management Considerations

Implementations receiving set objects that they consider out of order MAY log this. That could be helpful for diagnosing backward compatibility issues.¶

8. Other Efforts

In the past there have been effort to consolidate and update the RBNF such as in [I-D.cmfg-pce-pcep-grammar]. This document document relaxes the object ordering only, it does not take on the various other issues or the need to consolidate the RBNF for all PCEP extensions. There have been proposal to consolidate the RBNF for the PCEP message in a single place in GitHub and use modern data modeling tools to represent PCEP extensions. They might be taken up in parallel.¶

9. Security Considerations

This document does not raise any security issues.¶

10. IANA Considerations

This document does not require any IANA actions.¶

Appendix A. Acknowledgments

Thanks to John Scudder for the motivation behind this document. Thanks to Oscar Gonzalez de Dios and Cyril Margaria for raising errata on this topic. Thanks to the author of [I-D.cmfg-pce-pcep-grammar] for highlighting the issue.¶

Appendix B. Examples

As described in [EID6627], for the PCReq message, the CLASSTYPE object is encoded after the END-POINTS object in [RFC5455]. Where as in [RFC8231], the LSP object is encoded just after the END-POINTS object. So it is not known which of the below order is expected.¶

...<END-POINTS>[<LSP>][<CLASSTYPE>]...

or

...<END-POINTS>[<CLASSTYPE>][<LSP>]...

This update require the receiver to be able to except both of these.¶

Appendix C. When Order Matters

There are cases where the ordering between objects is important. For instance PCRpt message [RFC8231] includes <path> with some attributes say BANDWIDTH can be part of both <actual-attribute-list> and <intended-attribute-list>.¶

    Where:
      <path>::= <intended-path>
                [<actual-attribute-list><actual-path>]
                <intended-attribute-list>

An important factor to distinguish between the actual and intended attribute list is the presence of RRO (i.e. <actual-path>) and the order of objects in the PCRpt message.¶

If the RRO is present, any attributes encoded before it, are to be considered as part of <actual-attribute-list> and those after it, as part of <intended-attribute-list>.¶

If the RRO is absent, all attributes are part of <intended-attribute-list>.¶

Thus the approach taken by this document is to say that ordering is relaxed in cases where there is no ambiguity.¶

Author's Address