Summary: Has a DISCUSS. Has enough positions to pass once DISCUSS positions are resolved.
I support Alvaro's DISCUSS point #2. I'm confused about what the registration policy is for metrics in the new registry. If it is Specification Required, then the places in the document that assume new metrics are defined in an RFC need to be generalized, because Specification Required need not involve any RFC at all. I have an additional concern about this text: "If the proposed registry entry is defined in an RFC but is not yet widely deployed, there SHOULD be a statement in the RFC that says the proposed registry entry is not ready for registration, and use SHOULD employ a private/experimental ID. It is the responsibility of the document authors to submit the request to IANA when the proposed registry entry is ready for official registration." This appears to put a requirement on RFCs to include language that is not timeless and may later become out of date. That is, if this guidance is followed but a metric is later widely deployed, the RFC would have to be updated just to remove the text about the metric not being ready for registration. It seems better to just give guidance about which identifier range registration requests should target, and to give guidance to the designated experts about how to evaluate requests in different ranges.
Section 1: "any other organization that wishes to create a Performance Metrics Registry" It seems like performance metrics registry should not be capitalized here since the next section defines the capitalized version as the one maintained by IANA. Section 6: I would strongly recommend that this section be moved to an appendix. Some readers may find it useful, but it doesn't seem like it belongs in the body of the document. Also, I would caution against a section heading entitled "Why this Attempt Will Succeed." I certainly hope it will, but the future is hard to predict and the IETF has a long history of being wrong about what will and will not succeed.
Please be aware of the following note in the IANA section (which might simply reviewing): "Mock-ups of the implementation of this set of requests have been prepared with IANA's help during development of this memo, and have been captured in the Proceedings of IPPM working group sessions. IANA is currently preparing a mock-up. A recent version is available here: http://encrypted.net/IETFMetricsRegistry-106.html"
* Section 9. Please consider adding something approximately along the lines of: The aggregated results of the performance metrics described in this registry can reveal network topology information that might be considered sensitive. If this is the case, access control mechanism should be applied. * Editorial Nits: - Section 7.1.2. Typo. s/obervations/ observations/ - Section 7.1.3. Typo. s/formated/formatted/ - Section 7.3.1. Typo. s/unambigious/unambiguous/ - Section 7.3.2. Typo. s/wether/whether/ - Section 7.3.5. Typo. s/accomodated/accommodated/ - Section 7.4.4. Typo. s/calbration/calibration/ - Section 8.1. Spacing. s/[RFC8126]policy/[RFC8126] policy/ - Section 8.2. Typo. s/Regsirty/Registry/ - Section 8.2. Typo. s/unchamged/unchanged/ - Section 9. Typo. s/secuity/security/
The sometimes-conversational tone of the text here is at odds with my mental preconception of what a BCP (or, for that matter, a Standards-Track document) is, though I make no claim that my mental preconception is relevant to anyone other than myself. I agree with the comment about parts of the text assuing that metric specifications will be RFCs being at odds with the lower specification-required regisration policy. I support Alissa's Discuss point. Section 1 applications transported over its protocols. Performance metrics are such an important part of the operations of IETF protocols that [RFC6390] specifies guidelines for their development. It's not entirely clear that the existence of RFC 6390 supports exactly this conclusion. The definition and use of Performance Metrics in the IETF happens in various working groups (WG), most notably: Is this going to age well? The "Performance Metrics for Other Layers" (PMOL) a concluded WG, nits: s/a //, no comma Section 2 Examples of Performance Metrics are the FTP response time for a complete file download, the DNS response time to resolve the IP address, a database logging time, etc. This definition is nit: DNS responses are not limited to single IP addresses, so the definite article seems wrong here. consistent with the definition of metric in [RFC2330] and broader than the definition of performance metric in [RFC6390]. Why are we using a different definition of performance metric than BCP 170? [How can both be BCPs at the same time but be in conflict?] Parameter: A Parameter is an input factor defined as a variable in the definition of a Performance Metric. A Parameter is a numerical or other specified factor forming one of a set that defines a metric or sets the conditions of its operation. All Parameters must be known to measure using a metric and interpret the results. There are two types of Parameters: Fixed and Run- nit: I suggest "all Parameters must be known in order to make a measurement using a metric", as currently "known to measure" flows oddly. Section 3 groups: IPPM, XRBLOCK, IPFIX, and BMWG. This document analyzes an prior attempt to set up a Performance Metrics Registry, and the nit: s/ an/ a/ Based on [RFC8126] Section 4.3, this document is processed as Best Current Practice (BCP) [RFC2026]. As was already noted, the referenced section does not have any obvious relevance to publication status for this document. Section 4.1 As any IETF registry, the primary use for a registry is to manage a nit: I think this was intended to be "As for any IETF registry" particular example is the LMAP framework [RFC7594]. Using the LMAP terminology, the Performance Metrics Registry is used in the LMAP Control protocol to allow a Controller to request a measurement task to one or more Measurement Agents. In order to nit: please check the grammar of "request a measurement task to". Section 4.2 different (and incompatible) ways. Having a registry would allow both the IETF community and external people to have a single list of nit: I suggest rephrasing "external people" as needlessly divisive; perhaps "other communities"? Section 7 form of registry extension. The specification defining the new column(s) MUST give guidelines to populate the new column(s) for existing entries (in general). Is "MUST (in general)" really still a "MUST"? Section 7.1.2 RTDelay (Round Trip Delay) RTDNS (Response Time Domain Name Service) It's slightly unfortunate that "RT" expands to both "Round Trip" and "Response Time". o SubTypeMethod: One or more sub-types to further describe the features of the entry, such as: ICMP (Internet Control Message Protocol) IP (Internet Protocol) DSCPxx (where xx is replaced by a Diffserv code point) I thought Diffserv code point interpretation was a local matter; is it really appropriate in the metric definition? Section 7.3.5 devices. For example, parameters that include an IPv4 address can be encoded as a 32 bit integer (i.e. binary base64 encoded value) or ip- address as defined in [RFC6991]. The actual encoding(s) used must be nit: YANG types are specified distinctly from their possible encodings; indicating an XML or JSON encoded version of the YANG ip-address would be appropriate here. explicitly defined for each Run-time parameter. IPv6 addresses and options MUST be accomodated, allowing Registered Metrics to be used in either address family. nit: is "either" too limiting, here? Section 8.2 A change to a Registered Performance Metric SHALL be determined to be backward-compatible only when: Is it better to give the Experts some leeway to do the right thing rather than locking it down with rigid procedures? Section 8.3 The use of deprecated Registered Performance Metrics should result in a log entry or human-readable warning by the respective application. Taken literally this would require any measurement agent to query IANA before performing every measurement, to confirm that the metric being used is not deprecated. Presumably some less-frequent polling frequencey would still be acceptable! Section 10.2 reviewed along with the metric entry. New assignments for Registered Performance Metric Name Elements will be administered by IANA through Expert Review [RFC8126], i.e., review by one of a group of experts, the Performance Metric Experts, who are appointed by the IESG upon recommendation of the Transport Area Directors. It's not entirely clear to me that we need to go into this much detail when RFC 8126 covers it already. (Also elsewhere.) Section 11.1.3 Please use https:// URIs. Section 13.2 One might argue that RFC 7799 should be normative, since we defer to it for the allowed valuse for the Method column.
Why, oh why didn't I read this before draft-ietf-ippm-initial-registry? 'twould have made things much simpler :-)
I support Álvaro’s second DISCUSS point and the related issues he lists after it. I also think we should resolve Álvaro’s first DISCUSS point by simply putting a note in the IANA Considerations in this document that refers to the other document as providing initial registrations, and not waste time and effort doing a formal expert review of registrations that the working group has already reviewed and accepted here.
Thanks for addressing my DISCUSS.
Thank you for this document. I only have a couple of comment/questions, and found few nits while reading. Performance Metrics Registry MAY use the same I'm not sure 2119/8174 language is needed here. I'm not sure to understand the Version (of Registry Format) column. All entries will have the same number there, right? If this RFC-to-be is updated I guess we'll change to version 2. What shall happen to existing entries, will they keep Version=1 or adopt Version=2? s/in order to included/in order to be included/ As any IETF registry, the primary use for a registry is to manage a registry for its use within one or more protocols. This sentence seems a bit hard to parse s/other form of Performance Metric/other form of Performance Measurement/ ?
Thank you for the work put into this document. It is quite extensive (I even wonder whether it was useful to indicate the motivations for the registry). There are a couple of COMMENTs below; feel free to ignore them but I would appreciate if you replied. Regards, -éric == COMMENTS == -- Section 5 -- "interpretable by the user." who is the user in this case? (I have my guess of course but let's try to be clear) Why specifying "implementable by the **software** designer," ? I.e., are HW designers out of scope ? "accurate" is also quite vague Also a couple of nits in the section about ',' or '.' and upper/lower case characters. -- Section 6.1 -- s/will/should/ in "Why this Attempt Will Succeed" ? ;-) -- Section 7 -- The table part below is quite unclear at first and second reading. Worth re-wording ? " Category ------------------ Column | Column | " Or perhaps use a tree form (à la YANG module tree) ? -- Section 7.1..2 -- Probably worth mentioning "such as and not limited to" rather than "such as" ? It is also unclear how the MetricType, Method, ... can be extended. -- Section 11.1.3. -- Should the URL be an https:// one ?