Hi Luigi, Thanks for your response.
On the last point (the one that still needs some clarification): <snip> >> 2.- draft-li-nsa-reliability-00 >> ******************************* >> >> - In section 3, the draft states: >> >> "In the specific case of NSA >> [I-D.li-6lo-native-short-address], all nodes, except the root, have >> at least one secondary parent" >> >> Is this a feature of NSA enforced by design? > If talking about "parent" it is because the root by definition has no > parent node. > However, your comment make us realize that we did not emphasize enough > that for reliability there are at least two root nodes. > Thanks. > We will improve the language in the next revision so to make clear the > reliability from the root node perspective. Thanks, this will be useful. >> It might be helpful to discuss in the document(s) how scenarios >> without a secondary parent are supported. For example, a string >> topology, such as: >> >> A ----- B ----- C ----- D >> > > I would argue that in such topology there is no reliability ;-) > > Let me try to understand better your point. > Do you mean a "logical" strip topology? Implying that the underlaying > physical topology has redundant links? Well, my point was rather, going back to this sentence: "all nodes, except the root, have at least one secondary parent" - If "secondary parent" here means "secondary root", then everything is clear to me. - If "secondary parent" here means that a node must have at least two direct neighbors that can be used as parents (one being the primary one, and then a back-up one), then I was wondering whether this was enforced somehow and thus would exclude a network topology such as the one I referred to above, such as: A ---- B ---- C ---- D. Thanks, Carles (as a WG participant) > Thanks > > Ciao > > L. _______________________________________________ 6lo mailing list 6lo@ietf.org https://www.ietf.org/mailman/listinfo/6lo
