Hi Tony, On Apr 18, 2014, at 6:28 PM, prz <[email protected]> wrote:
> On Fri, 18 Apr 2014 20:41:13 +0000, Acee Lindem <[email protected]> > wrote: >> Hi Peter, >> >> On Apr 18, 2014, at 12:57 PM, Peter Psenak <[email protected]> wrote: >> >>> Hi Acee, >>> >>> I do not see any problem with routers with various compatibility modes >>> living in area/network together. With all the modes we have, we provide the >>> possibility of graceful migration from old to new LSA types. I don't think >>> we need to restrict any combination. Sure, if someone mix "full" and >>> "none", then it can not expect the full reachability between them, but that >>> does not mean we have to prohibit it. >> > > Zipped through the diffs on the new one. See most work went into migration. > Two observations: > > . nitpick: WILL is _not_ any kind of normative language. Use MUST Ok - I’ll search these out. > . the only reason the migration holds together without stable loops is that > there is an implicit assumption in the draft (I didn't see it spelled out but > then again I skimmed over it only) that the metric advertised in TLVs and old > style is always absolutely the same. I suggest to spell that out strongly. It is implicit but it seems rather obvious that if one originates the non-extended and extended versions of the same LSA, the content MUST be the same. However, I can state this explicitly. Thanks, Acee > > --- tony _______________________________________________ OSPF mailing list [email protected] https://www.ietf.org/mailman/listinfo/ospf
