Reviewer: Matthew Bocci
Review result: Has Nits
Thanks for a clear an well written draft. I have reviewed this from the
perspective of my knowledge of PCEP and the way PCCs and PCEs generally work,
rather than going through the YANG with a fine-toothed comb, as I am not really
a YANG expert and I would hope that a YANG doctor's review would cover the
syntax and other correctness of the YANG. I have just a few nits/questions,
below, but otherwise I think the draft is ready for publication.
The line numbers are from the IDNits output.
[snip]
16 Abstract
18 This document defines a YANG data model for the management of Path
s / Path / the Path
19 Computation Element communications Protocol (PCEP) for communications
[snip]
91 1. Introduction
93 The Path Computation Element (PCE) defined in [RFC4655] is an entity
94 that is capable of computing a network path or route based on a
95 network graph, and applying computational constraints. A Path
96 Computation Client (PCC) may make requests to a PCE for paths to be
97 computed.
99 PCEP is the communication protocol between a PCC and PCE and is
100 defined in [RFC5440]. PCEP interactions include path computation
101 requests and path computation replies as well as notifications of
102 specific states related to the use of a PCE in the context of
103 Multiprotocol Label Switching (MPLS) and Generalized MPLS (GMPLS)
104 Traffic Engineering (TE). [RFC8231] specifies extensions to PCEP to
105 enable stateful control of MPLS TE LSPs.
107 This document defines a YANG [RFC7950] data model for the management
108 of PCEP speakers. It is important to establish a common data model
109 for how PCEP speakers are identified, configured, and monitored. The
110 data model includes configuration data and state data.
112 This document contains a specification of the PCEP YANG module,
113 "ietf-pcep" which provides the PCEP [RFC5440] data model. Further,
114 this document also includes the PCEP statistics YANG module "ietf-
115 pcep-stats" which provides statistics, counters and telemetry data.
117 The PCEP operational state is included in the same tree as the PCEP
118 configuration consistent with Network Management Datastore
119 Architecture (NMDA) [RFC8342]. The origin of the data is indicated
120 as per the origin metadata annotation.
MB> I take the above text to mean that this draft is a YANG model for PCEP when
the data plane is assumed to be MPLS. However, it doesn't quite say that. It
seems to imply that MPLS is the only valid data plane, when in fact SRv6 could
be used and there are drafts related to that. I would suggest rephrasing or
adding text to say the PCEP in general could be used with other data planes,
but we are only modelling MPLS here, or something along those lines. Just to
make it very clear what the scope of the model is.
[snip]
553 | +--rw inter-layer? boolean {inter-layer}?
554 | +--rw h-pce {h-pce}?
555 | +--rw enabled? boolean
556 | +--rw stateful? boolean {stateful}?
557 | +--rw role? hpce-role
558 +--rw msd? uint8 {sr}?
MB> The model implies that a PCC could have a MSD configured that is different
from the MSD that is advertised in the IGP, for example. I thought MSD was
really a router/LER-wide property, determined by the underlying datapath
implementation, rather than something to configure, so should this not be YANG
state for the PCC (i.e. ro rather than rw )? This question is also applicable
to line 771 in the draft.
_______________________________________________
Pce mailing list -- [email protected]
To unsubscribe send an email to [email protected]
