Hi Tony,

Yes, sorry, I had in mind the reverse approximation to the problem. Indicating 
current power, power-save capability and current power-save status (including 
interfaces) the resulting power-group hierarchy has enough information and you 
don’t need calculations for the sleeping bw but maybe I’m missing something 
obvious.

Thanks,

Raúl Arco.

De: Tony Li <[email protected]> en nombre de Tony Li <[email protected]>
Fecha: jueves, 31 de julio de 2025, 17:56
Para: Raul Arco (Nokia) <[email protected]>
CC: Ron Bonica <[email protected]>, lsr <[email protected]>
Asunto: Re: [Lsr] New Version Notification for draft-many-lsr-power-group-00.txt


CAUTION: This is an external email. Please be very careful when clicking links 
or opening attachments. See the URL nok.it/ext for additional information.



Hi Raul,


While I see interesting to announce that a power group is power-save capable, I 
would also see interesting the possibility to advertise that a power-group is 
in fact in power-save mode. Announcing the capability could be useful for 
external elements to later interact with the node but advertising the current 
power-save mode can be also useful for standalone node decisions like 
energy-aware flex-algo for example.


For our use case, we believe that it’s both necessary and sufficient for the 
interfaces in the power group to be put into power sleep, and hence withdrawn 
from neighbor advertisement. Once all members of a power group are withdrawn, 
the power group itself may also be withdrawn (or not).  We do not update the 
state of the power group because it does not affect path computation and 
churning the LSDB has a network wide impact.


About the unidirectional Sleeping BW, could you expand on the use case?.


Consider the case where there is a LAG of 10 100Gbps interfaces.  3 of them 
have been put into power sleep.  How do we inform others that there is 300Gbps 
available but not otherwise advertised at present?

T


_______________________________________________
Lsr mailing list -- [email protected]
To unsubscribe send an email to [email protected]

Reply via email to