Hi Henning,
thanks for your review, please see some responses inline (##PP):
On 01/11/2021 11:18, Henning Rogge via Datatracker wrote:
Reviewer: Henning Rogge
Review result: Ready
Hi,
I have been asked to do a review of draft-ietf-lsr-flex-algo (currently in
draft 18).
First I have to say its very interesting to see that the similarities for the
routing/forwarding backplanes make it possible to have a common RFC that
modifies routing behavior for multiple routing protocols.
My area of knowledge is coming from the MANET area, but I can clearly see that
the mechanisms of this draft are generic enough to apply to most
'multi-topology' capable routing protocols. I wonder if a similar approach
could also be used in MANETs to clearly define sub-topologies that can then be
used to forward trafic along.
I think the draft is well written, most issues I had with the document were
caused by me not knowing the referenced other documents (which would not be a
problem for someone implementing this draft).
The draft uses both OSPF and OSPFv2 and OSPFv3 in the text. I assume that every
text using "OSPF" means "both v2 and v3"?
##PP
yes, that is indeed the wording that we used in many past documents.
I have also a few questions about specific chapters...
In chapter 6.1-6.3 (7.1-7.3 for OSPF) you define a subset of "boolean algebra"
among the extended admin groups. I assume that more complex boolean conditions
are not necessary for most deployments (and could be done by defining more EAGs
that represent the specific condition)?
##PP
the include/exclude rules that we defined have been used in traditional
RSVP-TE for a long time. They have proved to be sufficient. We can
always add more if needed in the future.
Chapter 7.4 defines that the Flags field has to have a length with a multiple
of 4 octets, which is not the case for the definition in chapter 6.4. Is the
length constraint only necessary in OSPF?
##PP
yes, OSPF TLVs are always 4 bytes aligned, which is not the case for
ISIS TLVs.
The IANA registries all state "this document" as a reference (and most of them
add a link to a chapter). I think the "this document" could just be omitted to
shorten the reference.
sure, will remove "this document" with the next update.
thanks,
Peter
Henning Rogge
--
last-call mailing list
last-call@xxxxxxxx
https://www.ietf.org/mailman/listinfo/last-call
