> Stretch and microsleeps are solution space and I'd argue that the best > solutions for core may involve neither. There is no highly credible > evidence that either has an impact on core network energy usage and > IMO stronger arguments that it would have no effect or infeasible > (microsleeping finely tuned temperature compensated colored lasers?) > than the the very weak arguments (mostly speculation) that there is > benefits to these two approaches for core.
No, stretch and sleep states are not "in the solution space." If you're going to save energy by using sleep states, then you need to deal with state. These are bars any given design must overcome, not suggestions for a design that would work. > So the research happen where it belongs - elsewhere. Then when there > is credible research presented, it might be time to consider a > taxonomy of the problem and solution space in an SDO. So your basic argument is --"we should only consider this when it's proven to be economically viable." My counter to that is simple: There's more to this game than saving more than x$/year. There's also the impact on the control plane to consider as a "cost." This draft is an attempt to at least point those costs out in a reasonable and understandable way. In other words, it's trying to get all the requirements on the table, rather than focusing myopically on the economic costs. Again, it's answering a completely different question than you're asking. Russ _______________________________________________ rtgwg mailing list [email protected] https://www.ietf.org/mailman/listinfo/rtgwg
