Igor, On Nov 9, 2012, at 11:21 AM, Igor Bryskin wrote:
> Jan, > > You said: > => draft-ietf-pce-stateful-pce does not change the control of the LSPs - it > still remains with the PCC. A network controller would have to be able to > dynamically add/delete LSPs in order to fit in the model that you described > above. draft-crabbe-pce-pce-initiated-lsp-00 goes there, but > draft-ietf-pce-stateful-pce does not. > > By this you are acknowledging that statefull pce can do a sub-set of things > that stateless pce - network controller combo is able to do. So what is the > point to discuss a solution that is worse than existing one? > What we got here is... failure to communicate. draft-etf-pce-stateful-pce is about a *stateful* PCE than can make temporal and spatial decisions about LSP placements. You are trying to compare it to a *stateless* PCE connected to a controller. Note that a controller can not make decisions about when to instantiate a particular LSP - the PCE must make those decisions, if you want to address use cases outlined in draft-ietf-pce-stateful-pce. Note also that a stateless PCE does not have the knowledge about LSPs in the network, so it can't make the same optimization decisions as a stateful PCE which does have the knowledge about LSPs; we need LSP knowledge in the PCE to address use cases outlined in draft-ietf-pce-stateful-pce. I don't think that one solution is better than the other. They are different, because they are addressing different issues. > Igor Thanks, Jan _______________________________________________ Pce mailing list Pce@ietf.org https://www.ietf.org/mailman/listinfo/pce