|
17
|
1 https://datatracker.ietf.org/doc/draft-ietf-sidrops-avoid-rpki-state-in-bgp/reviewrequest/24139/login/
|
|
18
|
2 Document: draft-ietf-sidrops-avoid-rpki-state-in-bgp-11
|
|
|
3 Title: Guidance to Avoid Carrying RPKI Validation States in BGP Path Attributes
|
|
|
4 Reviewer: Henry S. Thompson
|
|
|
5 Review Date: 2025-06-17
|
|
|
6
|
|
|
7 *Summary*
|
|
|
8
|
|
|
9 As far as I can tell as a non-specialist, the presentation of this
|
|
|
10 proposed BCP is in good shape. There is one minor suggestion and some
|
|
|
11 editorial nits below.
|
|
|
12
|
|
|
13 *Minor point*
|
|
|
14
|
|
|
15 1.
|
|
|
16
|
|
|
17 ... "states leak through to EGBP speakers." Would it be possible to
|
|
|
18 _briefly_ mention here at least one known way in which such leakage
|
|
|
19 can occur in a conformant router? That way operators of 'good'
|
|
|
20 routers may realise this document is relevant to them.
|
|
|
21
|
|
|
22 *Nits*
|
|
|
23
|
|
|
24 1.
|
|
|
25
|
|
|
26 In this phrase
|
|
|
27
|
|
|
28 "Annotating ... Attribute carried across EBGP session (based on validation
|
|
|
29 states) means that BGP UPDATE messages have to be sent to every
|
|
|
30 neighbor when the aforementioned validation states"
|
|
|
31
|
|
|
32 there's a minor grammatical problem. I would suggest the following
|
|
|
33 change
|
|
|
34
|
|
|
35 "... Attribute carried across an EBGP session (based on validation
|
|
|
36 states) means that BGP UPDATE messages have to be sent to every
|
|
|
37 neighbor when the aforementioned validation states"
|
|
|
38
|
|
|
39 4.
|
|
|
40
|
|
|
41 "BGP communities" -> "BGP Communities" [here and in 4 other places below]
|
|
|
42
|
|
|
43 "hold similar" -> "hold in a similar way"
|
|
|
44
|
|
|
45 4.1
|
|
|
46
|
|
|
47 "a large amounts" -> "large amounts"
|
|
|
48
|
|
|
49 4.1.2
|
|
|
50
|
|
|
51 "caused by operators to signal validation states" ->
|
|
|
52 "caused by operators signalling validation states"
|
|
|
53
|
|
|
54
|
