[Idr] Re: AD evaluation review of draft-ietf-idr-link-bandwidth-15

[Robert Raszuk <[e-mail]>]

Anoop,

How you are load balancing traffic is fully dependent on the choice of
underlay and operator's preference.

Therefor I don't think subject draft is a right place to discuss those
aspects in detail.

The reason I brought up unequal cost lb was only to make sure text in the
draft is not limiting - nothing more nothing less. Ketan's last suggestion
seems as said spot on.

Best,
R.

On Fri, Sep 5, 2025 at 9:18 PM Anoop Ghanwani <[e-mail]> wrote:

> Is there an RFC that defines what weighted ECMP and/or UCMP are?
>
> Searching around on the net it looks like there are at least some vendor
> docs that use the term interchangeably (and explicitly state so).
>
> If there is a difference between the two and we want this draft to be used
> for both, then perhaps all instances of WECMP in the draft would need to be
> revisited for consistency.
>
> Thanks,
> Anoop
>
> On Fri, Sep 5, 2025 at 1:06 AM Robert Raszuk <[e-mail]> wrote:
>
>> Perfect on both points raised.
>>
>> On Fri, Sep 5, 2025 at 6:16 AM Ketan Talaulikar <[e-mail]>
>> wrote:
>>
>>> Hi Robert/Authors,
>>>
>>> Perhaps:
>>> This document specifies a type of BGP Extended Community that enables
>>> routers to perform
>>> weighted load-balancing in multipath scenarios.
>>>
>>> Thanks,
>>> Ketan
>>>
>>>
>>> On Fri, Sep 5, 2025 at 12:09 AM Robert Raszuk <[e-mail]> wrote:
>>>
>>>> Hi,
>>>>
>>>>
>>>>> 19 Abstract
>>>>>
>>>>> 21   This document describes an application of BGP extended communities
>>>>> 22   that allows a router to perform WECMP (Weighted Equal-Cost
>>>>> 23   Multipath).
>>>>>
>>>>> <major> The document is actually specifying a new BGP extended
>>>>> community.
>>>>>
>>>>> SUGGEST:
>>>>> This document specifies a BGP Extended Community that enables routers
>>>>> to perform
>>>>> weighted equal-cost multipath (WECMP).
>>>>>
>>>>
>>>> This document defines a new type not BGP Extended Community as an
>>>> attribute. I actually find the original text more correct.
>>>>
>>>> Actually while we are at this sentence I would like to point out that
>>>> it can be also useful to perform unequal-cost multipath.
>>>>
>>>> Especially with encapsulation and EPE the cost to more then one egress
>>>> does not need to be equal for multipath. That's legacy for hop by hop IP
>>>> lookup networks which is rather gone these days.
>>>>
>>>> Thx,
>>>> Robert
>>>>
>>> _______________________________________________
>> Idr mailing list -- [e-mail]
>> To unsubscribe send an email to [e-mail]
>>
>