Interesting observation indeed.
I guess it depends on macro view if we are talking global or intradomain
("limited domains").
Cheers,
R.
On Sun, Nov 26, 2023 at 6:05 PM David Farmer <[email protected]> wrote:
> I don’t think you need multi path to utilize this anycast community,
> neither do you need to override hot potato routing. In fact in my mind,
> when this community is set you want to ensure hot potato routing, even if
> your normal policy was not to use hot potato routing.
>
> For example, in the research and education (R&E) community, it is common
> to local preference routes from other R&E networks above the commodity
> Internet. However, routes marked with this anycast community would likely
> be an exception to this policy helping to ensure hot potato routing for
> anycast prefixes, even when the normal policy is not to use hot potato
> routing.
>
> Thanks.
>
>
>
> On Sun, Nov 26, 2023 at 09:02 Robert Raszuk <[email protected]> wrote:
>
>> Hi,
>>
>> So I reread this discussion and have one more question .. .
>>
>> Abstract says:
>>
>>
>>
>>
>> *To allow operators to take well informed decisions on which prefixesare
>> carrying anycast traffic this document proposes a well-known BGPcommunity
>> to denote this property.*
>>
>> Cool - but how is this decision going to be instantiated in the router ?
>>
>> Effectively to make use of this community (provided all good marking and
>> good will of transits) you need to have some hooks in your local BGP
>> implementations to allow for multipath selection when at least one received
>> prefix is marked with such community.
>>
>> Of course if someone already enables ebgp or ibgp multipath for all
>> prefixes there is nothing to do here, but if you are generally just
>> selecting a single best path and doing hot potato routing I am missing how
>> any operator will (or can) actually use this community effectively.
>>
>> Cheers,
>> Robert
>>
>>
>> On Sun, Nov 26, 2023 at 2:50 PM Job Snijders <job=
>> [email protected]> wrote:
>>
>>> Dear Maximilian,
>>>
>>> On Sun, Nov 26, 2023 at 02:45:58PM +0100, Maximilian Wilhelm wrote:
>>> > Anno domini 2023 Maximilian Wilhelm scripsit:
>>> >
>>> > [...]
>>> > > Some time has past and I don't see any new comments, neither here nor
>>> > > to us authors.
>>> > >
>>> > > Hence I'd like to ask the chairs to start the WG Last Call.
>>> >
>>> > I believe this didn't happen so far, or did I miss it? :)
>>> >
>>> > So I'd like to ask to start the WG Last Call, or, if there is no
>>> > interest in getting this into an RFC, to call that out. :)
>>>
>>> Thanks for the poke
>>>
>>> Looking back I see two messages with comments that probably need
>>> addressing:
>>>
>>> https://mailarchive.ietf.org/arch/msg/grow/5iJ-QpjcPXOwIGhuOQMKn_FYQQc/
>>> https://mailarchive.ietf.org/arch/msg/grow/cdlMR2A0-gHqDwIf-UNkBvqVixw/
>>>
>>> Kind regards,
>>>
>>> Job
>>>
>>> _______________________________________________
>>> GROW mailing list
>>> [email protected]
>>> https://www.ietf.org/mailman/listinfo/grow
>>>
>> _______________________________________________
>> GROW mailing list
>> [email protected]
>> https://www.ietf.org/mailman/listinfo/grow
>>
>
_______________________________________________
GROW mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/grow
