Hi all,

I just wanted to follow up on an idea that arose yesterday during discussion of Greg's management presentation.

I think that developing a standard protocol independent timing management entity would be a really good thing to do.

In particular, such an information base could hold more detailed information than simply

"PRC", or "stratum 2" or "wristwatch", and could be used by possible clients to make more informed

decisions as to which server to pick.

Such a MIB could hold the following kinds of information :

identification (name/ID)

types of timing supported (none, frequency, uncalibrated time, ToD)

supported timing functionality (grandmaster, server, TC, BC, …)

to what this timing is traceable (local hydrogen maser, GPS, …, sundial)

history of lock (hourly ?)

supported protocols (SyncE, NTPv3, NTPv4, 1588v1, 1588v2, CTP, TTP, sercos, …)

mean delay to upstream master, server, or peer

measured asymmetries

current UTC offset

dispersion over the past few days

port states, time domains, …

etc.

and would have traps for loss of lock events.

There could also be protocol dependent parts, holding internals that may be needed.

Comments ?

Y(J)S