[Idr] Re: WG Adoption call for draft-hegdearavind-idr-bgp-ls-flex-algo-ext, ends 31 October, 2025.

["Aravind Babu MahendraBabu (aramahen)" <[e-mail]>]

Hi John,

Thanks for the review. Please see inline.
Would it be fine if I can incorporate the changes after the WG adoption.

Thanks,
Aravind

From: John Scudder <[e-mail]>
Date: Monday, 27 October 2025 at 7:35 PM
To: Jeffrey Haas <[e-mail]>
Cc: idr <[e-mail]>, [e-mail] <[e-mail]>
Subject: Re: [Idr] WG Adoption call for draft-hegdearavind-idr-bgp-ls-flex-algo-ext, ends 31 October, 2025.
Hi All,

> On Oct 17, 2025, at 11:55 AM, Jeffrey Haas <[e-mail]> wrote:
>
> Please reply to this thread to indicate your support, or lack thereof, for adoption.

I’ve reviewed this draft and I think we should adopt it. It’s well-written and useful, and I think it does not need much more work before it's ready for WGLC.

I have a few minor comments, below.

—John

Comments:

- IMO you need a normative reference for IEEE floating point format. It looks like you can copy this from RFC 9843.
<Aravind> You have mentioned it in previous review as well. Will add this change.


- "Maximum Link Delay: The maximum link delay is encoded in microseconds”. Probably you should say that it’s an integer (especially since a bunch of other similar fields are IEEE floating points). Up to you if you want to say the traditional “in network byte order” part…
<Aravind> Agreed and will add.


- In the Maximum Link Delay Sub-TLV, it seems a little annoying that the Reserved field is after the Maximum Link Delay field and not before. If it were before, since it's MBZ, it would be cheaper to decode since you wouldn’t need a shift. Depending on implementation status or other factors that may have driven this decision, this ship may have sailed, but I thought I’d mention it. It kind of stuck out since the rest of the TLVs appear to have used padding carefully to align their fields (probably because of their IGP heritage).
<Aravind> There are some implementations in this way. Will check with co-authors and make required changes to the draft.


- It seems to me that instead of using “TBA” as a placeholder for ten different things, it would be better to use “TBA0 … TBA9”. This is presumably helpful to the RFCEd and IANA, and just as important, reduces the risk of an error being introduced when the code points are assigned.
<Aravind> Agreed and will update.


Nits:

- You write “e.g., OSPF and IS-IS” but that list isn’t just exemplary, but rather complete, isn’t it? (If one understands “OSPF” to mean both v2 and v3.) So, you could just remove “e.g.” in my opinion. (2x)

- "The FAD Exclude Minimum Bandwidth sub-TLV is an optional sub-TLV that is used to carry the minimum bandwidth associated with the FAD that are used in the computation”… There appears to be a disagreement in number, shouldn’t it be “that *is* used”, not “that are used”? Or, “the FADs that are used…”?

- Similar to the previous, "The FAD Bandwidth Thresholds sub-TLV is an optional sub-TLV that is used to carry the information needed for bandwidth threshold method of metric calculation associated with the FAD that are used in the computation of the specific algorithm”. (Also, you’re missing a definite article, shouldn’t it be “… the bandwidth threshold method…”?

- "Min Bandwidth” — since the diagram writes out “Minimum Bandwidth” in full I see no reason to abbreviate to “Min” in the description.
<Aravind> Will update all the nits comments in the next update.