Ryan wrote: > I suppose I'm just bothered - in an intellectual, obsessive-compulsive > way - by all the DSL-based pool servers out there advertising > stratum-2 but serving up time with multi-millisecond phase errors. > Especially when those errors are easy to calculate and correct. > > As mentioned before, it doesn't matter much for the accuracy of pool > clients, which have phase errors on the order of tens of milliseconds > in many cases. But perhaps we should suggest those on slower DSL > connections fudge the stratum of their servers down to "announce" this > systemic error in some way, so that nobody tries to use one of the > DSL-based pool servers for a more sensitive application by "accident".
But stratum number has little to do with supposed accuracy or precision. It has everything to do with preventing self-referential loops among servers. Fudging your own stratum to a number higher than it would otherwise be prevents loops, BUT currently stock ntpd only allows a server to fudge a fixed stratum number, not a stratum that is (say) 3 numbers higher than otherwise computed. So advertising a fixed stratum could conceivably lead to loops (although I'll agree its unlikely if the advertised stratum is simply a large number.) ntpd itself will weed out falsetickers just fine if necessary. Tim. _______________________________________________ timekeepers mailing list [email protected] https://fortytwo.ch/mailman/cgi-bin/listinfo/timekeepers
