Hi Tony, Thank you for the clarification. Please correct me if I am wrong, it seems you mean that the power-group is an abstraction of the GREEN component-level model, designed for efficient IGP encoding.
However, this reasoning reverses the logical order. Abstraction is only valid after the underlying model is stable and its semantics are well understood. The GREEN framework and relevant power state model — component granularity, hierarchy, state definitions, capabilities—is still under active discussion. If we define the power-group abstraction now, we are making design decisions about what to abstract, how to group components, and which dependencies to encode before the thing we are abstracting is settled. IGP encoding efficiency is a valid concern—after we know exactly what to encode. Adopting the drafts now risks freezing an abstraction that may not align with the final GREEN framework and the models, leading to the very integration pain we are trying to avoid. BR, Genchen From: Tony Li <[email protected]> Sent: 2026年9月2日 23:37 To: Chengen(GenChen) <[email protected]> Cc: Joel Halpern <[email protected]>; Bonica, Ron <[email protected]>; [email protected]; [email protected]; [email protected] Subject: Re: [Green] Re: [Teas] Re: [Lsr] Re: WG adoption call for draft-many-lsr-power-group-03 (Ends 2026-08-30) Hi Chengen, In this case, however, conflicts already exist. The power-group abstraction used in draft-many-lsr-power-group and draft-many-teas-power steering has no defined relationship to the power states and hierarchy being specified in the GREEN Framework and YANG module. As noted earlier, sleep‑wake coordination procedures are deliberately left out of the routing drafts, yet those procedures will inevitably depend on the very state model that GREEN is defining. As previously discussed, the power group abstraction is not at all in conflict with the GREEN model. GREEN models things at the component level. Highly appropriate and potentially quite detailed. Including that level of detail in the IGP is burdensome, as we have limited space and are trying to be much more efficient in our encoding. Instead, the power-group is an abstraction of the components that can be place into power-sleep. This allows multiple components that are power-sleep capable and have only a single dependency to be encoded as a single sub-TLV. Regards, Tony
_______________________________________________ Lsr mailing list -- [email protected] To unsubscribe send an email to [email protected]
