More definitions or reuse standard definitions from Lamport or Mills

[wtoorop]

Davis Mills

[KTeichel]

Short versions, loosely following David Mills (from his paper On the Chronometry and Metrology of Computer Network Timescales and their Application to the NTP):

• Time, Timescale: Time is an abstraction which determines the ordering of events in a given frame of reference called timescale.
• Oscillator: An oscillator is a device that outputs trigger signals, which are (to a given tolerance level) equidistant with respect to a timescale.
• (Physical) Clock, Counter: A (physical) clock is a device composed of an oscillator and a counter, where the counter records a number of cycles since its initialization, with increments triggered by the oscillator. (Note: It could make sense to go further into how increments can be made, since this is potentially relevant to our issue of adjusting clocks via frequency...)
• Timestamp: The recorded value of a clock's counter at a given time t on a given timescale is called the timestamp of that clock at t.

[partim]

I very much like these. Makes me think we could perhaps widen the scope of the draft to be the authoritative reference for when other documents have to deal with time?

[wtoorop]

You mean a terminology draft for time (like the terminology RFC for DNS)?

[KTeichel]

I honestly think that would merit its own document. I would love to get involved.

-----"wtoorop" notifications@github.com schrieb: -----

======================= An: "aanchal4/draft-timestamp-implementation-advice" draft-timestamp-implementation-advice@noreply.github.com Von: "wtoorop" notifications@github.com Datum: 07.06.2018 12:08 Kopie: "KTeichel" kristof.teichel@ptb.de, "Assign" assign@noreply.github.com Betreff: Re: [aanchal4/draft-timestamp-implementation-advice] More definitions or reuse standard definitions from Lamport or Mills (#16)

You mean a terminology draft for time (like the terminology RFC for DNS)?

-- You are receiving this because you were assigned. Reply to this email directly or view it on GitHub: https://github.com/aanchal4/draft-timestamp-implementation-advice/issues/16#issuecomment-395367196

[aanchal4]

• Time, Timescale: Time is an abstraction which determines the ordering of events in a given frame of reference called timescale.
• Oscillator: An oscillator is a device that outputs trigger signals, which are (to a given tolerance level) equidistant with respect to a timescale.
• (Physical) Clock, Counter: A (physical) clock is a device composed of an oscillator and a counter, where the counter records a number of cycles since its initialization, with increments triggered by the oscillator.
• Wall Time: The recorded value of a clock's counter at a given time t on a given timescale is called the timestamp of that clock at t.

Protocols and applications can express time in several forms, depending on whether or not universal agreement is required about that point in time. This section focuses on the differences between absolute time and relative time.

Expressing Time: 1) Absolute time expresses an absolute point in time (e.g., June 13, 2018 at 1:32:09pm). For instance,“Unix Time” is seconds since midnight January 1st, 1970, while“Universal Coordinated Time” (UTC) is an international time scale that forms the basis for the coordinated dissemination of standard frequencies and time signals [1]. Absolute time is often used to express the validity of objects with a limited lifetime that are shared over the network. In order to validate absolute time value, a system needs access to a reasonably close reference time, for instance one based on the UTC.

2) Relative time measures the time interval that has elapsed from some reference point (e.g., “20 minutes from the time of your query”). Relative time is commonly used in network protocols, e.g., to determine when a packet should be considered “dropped”, or e.g., to set Time To Live (TTL) values that govern the length of time for which an object is valid or usable. Relative time does not require access to the UTC time, or any other absolute time metric—only the rate of passage of his time across different systems is important.

References: [1] https://www.bipm.org/en/bipm-services/timescales/time-server.html

[aanchal4]

I have added two more definitions to that of Kristoff's.