Early Review of draft-ietf-sidrops-rtr-yang-05
review-ietf-sidrops-rtr-yang-05-yangdoctors-early-andersson-2026-05-27-00
| Request | Review of | draft-ietf-sidrops-rtr-yang |
|---|---|---|
| Requested revision | No specific revision (document currently at 07) | |
| Type | Early Review | |
| Team | YANG Doctors (yangdoctors) | |
| Deadline | 2026-05-04 | |
| Requested | 2026-04-10 | |
| Requested by | Luigi Iannone | |
| Authors | Yisong Liu , Changwang Lin , Haibo Wang , Jishnu Roy , Jeff Haas , Hongwei Liu , Di Ma | |
| I-D last updated | 2026-06-17 (Latest revision 2026-06-17) | |
| Completed reviews |
Yangdoctors Early review of -05
by Per Andersson
(diff)
|
|
| Comments |
This document proposes a YANG model, hence we would like to have an early review from the YANG doctors before going to WGLC. Thanks |
|
| Assignment | Reviewer | Per Andersson |
| State | Completed | |
| Request | Early review on draft-ietf-sidrops-rtr-yang by YANG Doctors Assigned | |
| Posted at | https://mailarchive.ietf.org/arch/msg/yang-doctors/Hr70NKyLZt13E4pgBxiLhZJzmxg | |
| Reviewed revision | 05 (document currently at 07) | |
| Result | Ready w/issues | |
| Completed | 2026-05-27 |
review-ietf-sidrops-rtr-yang-05-yangdoctors-early-andersson-2026-05-27-00
Hi! This is my Early Review of draft-ietf-sidrops-rtr-yang-05. My conclusion is that the YANG modules are in good shape but still need some work. Note that I am no expert in RTR, so take my comments as such. I am mainly reviewing the YANG modules in this document. Result: Ready with issues Review of the ietf-rpki-rtr YANG module --------------------------------------- Major issues: Section 3.2 need to say that it includes normative references to the following RFCs 9911, 8349, 7317, 9644, 8343, 8177, 6810, 8210, and probably 8340 as well. The leaf last-session-up-down is of type yang:timestamp, which is measured in hundreths of a second. However, the description states that the value is a timestamp in microseconds relative to the Unix Epoch. Suggest to change the type to uinte32 and either add 'unit "usec"' statement or present the time unit in the description, or both. The "unit" and description statements need to include the time unit for all of the timestamp and time leafs. Nit: Suggest to remove the -timestamp suffix used, it is possible infer that it is a timestamp from the rest of the name. Medium issues: The tree diagram in Sections 3.1 is not very helpful as it is are presented without context. I suggest that you take parts of the YANG trees and present them with some context, at a minimum references to relevant sections in e.g. RFC 8340. Minor issues: I prefer to present "config" statements early, especially before enumeration types. See e.g. the session-state leaf, I suggest to start with the "config false" statement. (This is the statement ordering used in the connection-data container.) The threshold-percentage leaf can use the rt-types:percentage type, I leave this decision to you. Review of the ietf-rpki-table YANG module ----------------------------------------- Major issues: Section 4.2 needs to state that it inludes normative references to the following RFCs 9911, 8349, 6487, 8608, and probably 8340 as well. Note that it is not explained anywhere what RFC YYYY is, guessing that it is draft-ietf-sidrops-rpki-ccr. Include this in the list of normative references as well, and either add a note that RFC YYYY is this draft, or update mentions of RFC YYYY to actually be this draft. Medium issues: The tree diagram could be broken up per component and explained with some context as suggested for ietf-rpki-rtr. Rename the ipv4-pfx-len and ipv6-pfx-len derived types to ipv4-prefix-length and ipv6-prefix-length. All other types have long names, no need to trade these bytes for lesser readability. If the RTR protocol doesn't require it to say "max-len" type it out in full "max-length" as well. Is it possible to tighten up the type for the router-key list leaf "key", which is a subjectPublicKeyInfo? This is outside my area of expertise, but it seems there are requirements on format and content? Nit: Suggest to remove the -time suffix used, it is possible infer that it is time from the rest of the name. General Nits: Section 2: OLD: Two YANG data models are defined in this document. NEW: Two YANG modules are defined in this document. OLD: The ietf-rpki-rtr.yang data model NEW: The ietf-rpki-rtr YANG module OLD: The ietf-rpki-table.yang data model NEW: The ietf-rpki-table YANG module $ pyang -f yang --yang-line-length=72 --ietf \ ietf-rpki-rtr@2026-03-31.yang > new.yang $ diff ietf-rpki-rtr@2026-03-31.yang new.yang 2c2 < yang-version "1.1"; --- > yang-version 1.1; 4c4,5 < prefix "rpki-rtr"; --- > prefix rpki-rtr; > 6c7 < prefix "yang"; --- > prefix yang; 11c12 < prefix "inet"; --- > prefix inet; 22c23 < prefix "ianach"; --- > prefix ianach; 27c28 < prefix "ssh"; --- > prefix ssh; 32c33 < prefix "if"; --- > prefix if; 45d45 < 64d63 < 97a97 > 104a105 > 111a113 > 114c116 < length 20; --- > length "20"; 118a121 > 123a127 > 177a182 > $ pyang -f yang --yang-line-length=72 --ietf \ ietf-rpki-table@2026-03-31.yang > new.yang $ diff ietf-rpki-table@2026-03-31.yang new.yang 2c2 < yang-version "1.1"; --- > yang-version 1.1; 4c4,5 < prefix "rpki-table"; --- > prefix rpki-table; > 6c7 < prefix "yang"; --- > prefix yang; 11c12 < prefix "inet"; --- > prefix inet; 25d25 < 44d43 < 78a78 > 85a86 > 92a94 > 95c97 < length 20; --- > length "20"; 99a102 > 102c105 < length 32; --- > length "32"; 117a121 > enum reconnect enum idle-forever s/Diconncets/Disconnects/ s/secruring/securing/ -- Per