2012/2/16 Thorsten Mühlfelder <[email protected]>: > Hello, > > I've already posted these questions 8 months ago without answers, but > now the problem came to my mind again. Sorry for double post. > > I'm using a Atmel uC embedded device as SNTP server. The code is > working so far, but I wonder what the correct settings are for: > * stratum > * precision > * root dispersion > * reference identifier > * reference timestamp > > The embedded device has a battery buffered hardware RTC, that is > capable of 1 second resolution. The used oscillator though has a > frequency tolerance of 20 ppm. > On startup the device tries to get the time from a pre-configured NTP > server once. > > In my understanding my server is a stratum 1 with reference identifier > LOCL. Also I chose reference timestamp as last time correction by > SNTP. Is this correct? > > I also wonder about the precision an root dispersion settings, as the > RTC only has a resolution of 1 second. > > Thanks for any help, I'd like to set it up the right way.
Your understanding is mistaken. The only case where it appropriate for a server not implementing the full NTP protocol to act as an NTP server is when it has a disciplined external reference to UTC, such as a local clock disciplined by other means to UTC time. The device you describe is retrieving time via NTP and must run the full protocol, which determines how to set the variables you mention, and does not allow for a single query at startup. For example, the stratum will be one higher than the upstream (or downstratum, if you prefer) NTP server. The root dispersion will be the sum of the source NTP server's root dispersion and a factor accounting for the effects of your link to that server. The precision would be 0, as it is a power of 2 seconds and 2**0 is one second. Precision -1 would indicate a clock which ticks at least twice per second but less than 4 times per second, etc. The reference timestamp would be the same as your source's. Please refer to RFC 5905 for all the details of NTP and SNTP. Cheers, Dave Hart _______________________________________________ questions mailing list [email protected] http://lists.ntp.org/listinfo/questions
