Export of Delay Performance Metrics in IP Flow Information eXport (IPFIX)
draft-ietf-opsawg-ipfix-on-path-telemetry-23
Yes
Mahesh Jethanandani
No Objection
Andy Newton
Erik Kline
Jim Guichard
Orie Steele
Paul Wouters
Note: This ballot was opened for revision 20 and is now closed.
Mahesh Jethanandani
Yes
Mohamed Boucadair
Yes
Comment
(2025-07-28 for -20)
Sent
Hi Thomas, Benoît, and Alex, Thank your for the effort put into this specification. Interesting to see the approach defined in RFC 8911 exercised here. Also, thanks to Menachem Dodge for the OPSDIR review and Qin Wu for the PERFMETRDIR review. Appreciate that Menachem, Qin, and authors converged for both reviews. As I reviewed a previous version of the spec [1], I only focused the current review on the diff since then [2]. I specifically appreciated the new text added to explain the need for the IEs compared to an approach where intermediate nodes export the timestamps observed in packets and local receiving timestamps and let the collector to do the various math operations. Likewise, I appreciate that the IEs were generalized to be independent of IOAM. I only have very minor comments: # Use the exact registry name ## Section 1 OLD: In order to export the delay-related metrics via IPFIX [RFC7011], this document defines four new IPFIX Information Elements (IEs), exposing the On-Path delay on OAM transit and decapsulating nodes, following the postcard mode principles. Since these IPFIX IEs are performance metrics [RFC8911], they are also registered in the "IANA Performance Metric Registry [IANA-PERF-METRIC]. NEW: In order to export the delay-related metrics via IPFIX [RFC7011], this document defines four new IPFIX Information Elements (IEs), exposing the On-Path delay on OAM transit and decapsulating nodes, following the postcard mode principles. Since these IPFIX IEs are performance metrics [RFC8911], they are also registered in the IANA “Performance Metric” Registry [IANA-PERF-METRIC]. ## Section 1 OLD: Following the guidelines for Registered Performance Metric Requesters and Reviewers [RFC8911], the different characteristics of the performance metrics (Identifier, Name, URI, Status, Requester, Revision, Revision Date, Description, etc.) must be clearly specified in the "IANA Performance Metric Registry [IANA-PERF-METRIC] in order NEW: Following the guidelines for Registered Performance Metric Requesters and Reviewers [RFC8911], the different characteristics of the performance metrics (Identifier, Name, URI, Status, Requester, Revision, Revision Date, Description, etc.) must be clearly specified in the IANA "Performance Metric” Registry [IANA-PERF-METRIC] in order ## Section 3.1 OLD: This section specifies four performance metrics for the Hybrid Type I Passive assessment of IP One-Way Delay, to be registered in the "IANA Performance Metric Registry [IANA-PERF-METRIC]. NEW: This section specifies four performance metrics for the Hybrid Type I Passive assessment of IP One-Way Delay, to be registered in the IANA “Performance Metric” Registry [IANA-PERF-METRIC]. # There might be multiple OAM encapsulating nodes OLD: Assuming time synchronization on devices, the delay is measured by calculating the difference between the timestamp imposed with On-Path Telemetry in the packet at the OAM encapsulating node and the ^^^^ NEW: Assuming time synchronization on devices, the delay is measured by calculating the difference between the timestamp imposed with On-Path Telemetry in the packet at an OAM encapsulating node and the Please check other similar constructs. # I don’t remember I have seen “flow packet” used in other IPFIX specs + there might be many OAM headers RFC7011 defines flow as follows: A Flow is defined as a set of packets or frames passing an Observation Point in the network during a certain time interval. All packets belonging to a particular Flow have a set of common properties. I suggest we make this change: OLD: * Encapsulating Node: Receives the IP Flow packets, encapsulates the ^^^^^^^^^^^^ packets with the OAM header and adds the timestamp into the OAM header. * Transit Node: Receives the IP Flow packets, measures the delay^ ^^^^^^^^^^^^ between the timestamp in the packet and the timestamp when the packet was received. * Decapsulating Node: Receives the IP Flow packets, computes the ^^^^^^^^^^^^ delay between the timestamp in the packet and the timestamp when the packet was received and removes the OAM header from the packet. NEW: * Encapsulating Node: Receives the IP packets, encapsulates the packets with an OAM header and adds the timestamp into the OAM ^^ header. * Transit Node: Receives the IP packets, measures the delay between the timestamp in the packet and the timestamp when the packet was received. * Decapsulating Node: Receives the IP Flow packets, computes the delay between the timestamp in the packet and the timestamp when the packet was received and removes the OAM header from the packet. # Please add “Collector” and “Observation Point” to the list of terms under 7011 CURRENT: The following terms are used as defined in [RFC7011]: … # (Nit) Only one term is imported from 7799 OLD: The following terms are used as defined in Section 3.8 of [RFC7799]: * Hybrid Type I Passive NEW: The following term is used as defined in Section 3.8 of [RFC7799]: * Hybrid Type I Passive # (nit) Section 3.1.2 OLD: We consider the measurement of one-way delay based on a single Observation Point [RFC7011] somewhere in the network. NEW: The measurement of one-way delay is based on a single Observation Point [RFC7011] somewhere in the network. Please make the same change for other similar constructs in the document. # IANA is the authoritative reference for IEs, not RFC7012 & Use the exact registry names OLD: 5.2. IPFIX Entities This document requests IANA to register new IPFIX IEs (see table 3) under the "IPFIX Information Elements" registry [RFC7012] available at "IANA IP Flow Information Export (IPFIX) Entities Registry [IANA-IPFIX] and assign the following initial code points. NEW: 5.2. IPFIX Entities This document requests IANA to register new IPFIX IEs (see table 3) under the "IPFIX Information Elements" registry under the " IP Flow Information Export (IPFIX) Entities” registry group [IANA-IPFIX] and assign the following initial code points. # RFC-to-be There are several notes to the RFC editor with distinct patterns [RCC-to-be] vs. _RFC[RFC-to-be] (1) with RFC label 778 Reference: [RFC-to-be] (2) without RFC label … 781 OWDelay_HybridType1_Passive_IP_RFC[RFC-to-be]_Seconds_Mean in the 782 IANA Performance Metric Registry. Hope this won't be confusing for the RFC editor. # Section 6.5 ## nit s/Section 4.4.2.3 and 4.4.2.4 of [RFC9197]/ Sections 4.4.2.3 and 4.4.2.4 of [RFC9197] There are many such occurrences in the doc. ## Cite RFC8754 CURRENT: 919 be applied to the SRv6 Segment Routing Header. Hope this helps. Cheers, Med [1] https://mailarchive.ietf.org/arch/msg/opsawg/qRNGYGSasUsON8PUvpqrP86ojyo/ [2] https://author-tools.ietf.org/iddiff?url1=draft-ietf-opsawg-ipfix-on-path-telemetry-08&url2=draft-ietf-opsawg-ipfix-on-path-telemetry-20&difftype=--html
Andy Newton
No Objection
Deb Cooley
No Objection
Comment
(2025-08-01 for -20)
Sent
Thanks to Linda Dunbar for their secdir review. Section 7, para 4: While RFC 7011 does indeed recommend TLS/DTLS to protect sensitive or privacy information, it pre-dates TLS 1.3 and DTLS 1.3. Consider adding a note to use BCP195, perhaps this: 'Implementations SHOULD also refer to [BCP195] for additional details.'
Éric Vyncke
No Objection
Comment
(2025-08-05 for -20)
Sent
# Éric Vyncke, INT AD, comments for draft-ietf-opsawg-ipfix-on-path-telemetry-20 CC @evyncke Thank you for the work put into this document. Please find below some non-blocking COMMENT points/nits (replies would be appreciated even if only for my own education). I also second many Gunter's points (about "mean" and IANA). Special thanks to Giuseppe Fioccola for the shepherd's detailed write-up including the WG consensus and the justification of the intended status. Other thanks to Tim Wicinski, the Internet directorate reviewer (at my request), please consider this int-dir review: https://datatracker.ietf.org/doc/review-ietf-opsawg-ipfix-on-path-telemetry-20-intdir-telechat-wicinski-2025-08-02/ I hope that this review helps to improve the document, Regards, -éric ## COMMENTS (non-blocking) ### Abstract Abstract should be short of course but this one is too succinct IMHO: be specific about encapsulating/decapsulating node (i.e., OAM header encaps and not VPN encaps) + a reference to RFC 9232 would be useful. ### Section 1 The term 'delay' is ambiguous in the IETF context... is it end to end ? serialisation ? propagation ? Only the 2nd paragraph explains it. s/postcard mode, where the metrics are also exposed in transit nodes/postcard mode, where the metrics are also exposed *by* transit nodes/ ? Again `decapsulating nodes` suggest using "OAM header decapsulating nodes". Unsure whether the next text is useful in an introduction and is not also easy to parse `Following the guidelines for Registered ... even if they are performed using different implementations and in different networks.` To make a much nicer HTML rendering, suggest using the aasvg too to generate SVG graphics. It is worth a try especially if the I-D uses the Kramdown file format ;-) `and/or the sum of measured path delay` it is unclear why there is a "and/or" and what is the "sum" here ? The accumulated delay by all segments or the sum of all delays locally computed ? If the latter, does it have any statistical value in the absence of sample count ? ### Section 3.3.2 [I-D.zhou-ippm-enhanced-alternate-marking] and [I-D.fz-spring-srv6-alt-mark] are clearly *normative* references. ### Section 3.3.5 Is the `host A Role` defined somewhere ? Add a forward reference ? or why not using "initiator" ? It seems to overly add some complexity. ### Section 3.3.6 The terms should rather be defined in the terminology section. ### Section 6.2 In 2025, does a division have a significant CPU impact when compared to other factors ? ### Section 6.4 Who is the "we" in `we might miss the temporal distribution` ? The authors ? The WG ? The IETF ? Please avoid using "we" in an I-D. ### Section 6.5 s/IPv6 options header/IPv6 *extensions* header/ ### Appendix A THANKS for the IPv6 examples ;-) s/us/µs/ as UTF-8 encoding is supported in I-D.
Erik Kline
No Objection
Gorry Fairhurst
No Objection
Comment
(2025-08-04 for -20)
Sent
Thanks to Martin Duke for his TSV-ART review, where he notes: This document could use an editorial review for clarity. Sentence fragments like "The timestamp when the packet is being received at OAM encapsulating node." are hard to parse. The Introduction was really unclear and I had to read it a few times to understand what this document was doing, despite a basic conceptual familiarity with IOAM. Please also see the DISCUSS comments of Gunter Van de Velde regarding operation of Alt Mark.
Gunter Van de Velde
(was Discuss)
No Objection
Comment
(2025-09-05 for -20)
Sent for earlier
Thank you for resolving all my discuss observations and taking my comments into consideration
Jim Guichard
No Objection
Ketan Talaulikar
No Objection
Comment
(2025-08-21 for -20)
Not sent
Thanks to the authors and the WG for their work on this document. I extend support to Gunter's DISCUSS position.
Mike Bishop
No Objection
Comment
(2025-08-05 for -20)
Sent
https://www.ietf.org/archive/id/draft-ietf-opsawg-ipfix-on-path-telemetry-20.html#section-1-5 feels like an argument for why the metrics are registered and why they're named what they are. Does the document need this, or should it just specify and register them? "Hybrid Type I Passive" is used as a single term in this document, saying it's imported from RFC7799. However, https://www.rfc-editor.org/rfc/rfc7799#section-3.8 appears to use "Hybrid Type I" and "Passive" as two separate terms, and RFC 7799 does not appear to contain the string "Hybrid Type I Passive". ===NITS FOLLOW=== - In 3.3.2, "Section 4.4.2.3 and 4.4.2.4" => "Sections 4.4.2.3 and 4.4.2.4", and these should probably be links to the sections in question. - In 3.3.5, should "hybrid type I" be capitalized as it is elsewhere in the document? - In the Acknowledgements, "Rest in Peace Al" => "Rest in Peace, Al" or simply "Rest in Peace" to avoid misreading this as a nickname.
Orie Steele
No Objection
Paul Wouters
No Objection