[Sea-time] Existing Time Management BoF charter & supporting material

Shames, Peter M (312B) Peter.M.Shames at jpl.nasa.gov
Tue Nov 6 21:31:09 UTC 2018


Hi Lee,

Helps a lot.  Thanks for the contribution.  I like the "operational domain" idea, that had occured to me as well.  The Internet analogy is a local / organizational NTP time server.

Did you guys also explore the different ways to handle clock synchronization on-board?

Thanks, Peter

PS – if you can share the "treatise" please do so on the website.



From: "Pitts, Robert L. (MSFC-HP27)[HOSC SERVICES CONTRACT]" <robert.l.pitts at nasa.gov>
Date: Tuesday, November 6, 2018 at 12:06 PM
To: Peter Shames <Peter.M.Shames at jpl.nasa.gov>, Jon Hamkins <Jon.Hamkins at jpl.nasa.gov>, "sea-time at mailman.ccsds.org" <sea-time at mailman.ccsds.org>
Subject: RE: [Sea-time] Existing Time Management BoF charter & supporting material

Actually I do have a treatise but I will spare you.

A few details,


1.       We did a survey several years ago to assess what level of accuracy and precision is required.  We found that the following accuracy/precision characterizes most missions.  These were mapped to different phase of mission where ultra and high were related to measurement activities and coarse to fine were navigational.  With navigational related to timescales.

Accuracy / Precision


  *   Coarse: Millisecond (msec) – 10-3 sec
  *   Fine: Microsecond (μsec) – 10-6 sec
  *   High: Nanosecond (ηsec) – 10-9 sec
  *   Ultra: Picosecond (psec) – 10-12 sec => (+20 years)



2.       So if you have a number of mission in the Mars arena, why not create a Mars time domain and offer services from infrastructure devices in orbit.  Similarly for the Lunar arena.  Services could be defined using multiple techniques.  However getting time to all devices within the domain need not be onerous and once in the domain, local relationships to include unique service or relativistic correction would be defined for all local mission.  The technique to set time in the domain would be standard and the services offered would subject to each mission.

3.       We also found that there is a need for multiple way to disseminate time.  Considerations include:

a.       One-way links

b.       Low bandwidth links

c.       Variable Latency

d.       Hop-by-hop transfers

e.       Periodic discontinuities

f.        Management of communication link loading

Does this help?

Lee

From: Shames, Peter M (312B) <Peter.M.Shames at jpl.nasa.gov>
Sent: Tuesday, November 06, 2018 12:45 PM
To: Pitts, Robert L. (MSFC-HP27)[HOSC SERVICES CONTRACT] <robert.l.pitts at nasa.gov>; Hamkins, Jon (JPL-3320)[Jet Propulsion Laboratory] <jon.hamkins at jpl.nasa.gov>; sea-time at mailman.ccsds.org
Subject: Re: [Sea-time] Existing Time Management BoF charter & supporting material

Thanks Lee.

For the benefit of all of us, could you possibly break down what you mean by "overall distribution and time domains in specific theatres of operation" to a more explicit set of concerns?  Something like a description of what you think "time distribution" and "time domains" mean, and also provide one or more examples of "specific theatres of operation"?  We don't need a treatise on the subject (unless you happen to have one at hand), but a little more detail will help.

Thanks, Peter


From: "Pitts, Robert L. (MSFC-HP27)[HOSC SERVICES CONTRACT]" <robert.l.pitts at nasa.gov<mailto:robert.l.pitts at nasa.gov>>
Date: Tuesday, November 6, 2018 at 10:28 AM
To: Peter Shames <Peter.M.Shames at jpl.nasa.gov<mailto:Peter.M.Shames at jpl.nasa.gov>>, Jon Hamkins <Jon.Hamkins at jpl.nasa.gov<mailto:Jon.Hamkins at jpl.nasa.gov>>, "sea-time at mailman.ccsds.org<mailto:sea-time at mailman.ccsds.org>" <sea-time at mailman.ccsds.org<mailto:sea-time at mailman.ccsds.org>>
Subject: RE: [Sea-time] Existing Time Management BoF charter & supporting material

Jon and Peter,

I think you are hitting on the areas that I am most concerned with to include overall distribution and time domains in specific theatres of operation.

Lee

From: SEA-TIME <sea-time-bounces at mailman.ccsds.org<mailto:sea-time-bounces at mailman.ccsds.org>> On Behalf Of Shames, Peter M (312B)
Sent: Tuesday, November 06, 2018 12:16 PM
To: Hamkins, Jon (JPL-3320)[Jet Propulsion Laboratory] <jon.hamkins at jpl.nasa.gov<mailto:jon.hamkins at jpl.nasa.gov>>; sea-time at mailman.ccsds.org<mailto:sea-time at mailman.ccsds.org>
Subject: Re: [Sea-time] Existing Time Management BoF charter & supporting material

Hi Jon,

Good questions.  As we have been addressing time synchronization that topic explicitly includes automated means doing this, especially in a multi-mission, networked (DTN), context.  I think of it as NTP for DTN.  This is a substantially bigger topic than a simple addition of an offset.

Do the rest of you agree?

Thanks, Peter


From: Jon Hamkins <Jon.Hamkins at jpl.caltech.edu<mailto:Jon.Hamkins at jpl.caltech.edu>>
Date: Tuesday, November 6, 2018 at 8:34 AM
To: Peter Shames <Peter.M.Shames at jpl.nasa.gov<mailto:Peter.M.Shames at jpl.nasa.gov>>, "sea-time at mailman.ccsds.org<mailto:sea-time at mailman.ccsds.org>" <sea-time at mailman.ccsds.org<mailto:sea-time at mailman.ccsds.org>>
Subject: Re: [Sea-time] Existing Time Management BoF charter & supporting material

Thanks, Peter.

The IOAG Service Catalog requests the following standards be developed:

• Clock Correlation Procedures [CC] - Magenta Book
• Time Transfer [TT] - Blue Book
• Time Synchronization [TS] - Blue Book

Do we all agree that these are still the appropriate books and colors to
be working toward?  Do we need to update our goals?  Thoughts?

On a separate topic, I have been thinking about what we discussed in
Berlin.  We mentioned time correlation and time synchronization as two
motivations for the work.  As we discussed, time correlation is
determining the offset between two clocks, and time synchronization is
setting two clocks so that they match.  The difference between these
concepts is quite small -- a time synchronized clock output is simply
the time correlated clock output time plus the value of the offset.  In
other words, after time correlation, a time offset is determined.
Whether this offset is used to explicitly synchronize a clock or is
simply added to the clock output each time it is read is an
implementation detail.  For this reason, I recommend that going forward,
we consider time correlation and time synchronization to be
fundamentally the same concept, with the same time correlation standard
governing them.

      ----Jon
On 11/5/2018 5:21 PM, Shames, Peter M (312B) wrote:
Dear Time BoF,
Attached is the current draft of the Time WG Charter and also the IOAG
Service catalog #2.  See pgs 26-27, section 4.5, to see what it is that
the IOAG has requested.  There is also an existing set of materials
relating to this topic that Ed Greenberg developed.
_______________________________________________
SEA-TIME mailing list
SEA-TIME at mailman.ccsds.org<mailto:SEA-TIME at mailman.ccsds.org>
https://mailman.ccsds.org/cgi-bin/mailman/listinfo/sea-time

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.ccsds.org/pipermail/sea-time/attachments/20181106/01e7a618/attachment.html>


More information about the SEA-TIME mailing list