[LEAPSECS] Time math libraries, UTC to TAI

[Steve Allen]

On Thu 2016-12-29T09:26:58 -0700, Warner Losh hath writ:
> How do you deal with acquiring knowledge of leap seconds after you've
> given out 'old' timestamps that might be affected by them? This is
> best described as how well you recover from the leapsecond file being
> corrupted and replaced, though there's many other scenarios here.

This question is similar to the ones that the EU financial sector
have been asking in the face of the ESMA MiFID II regulations that
will require extremly precise timestamps on market trades.
https://www.esma.europa.eu/sites/default/files/library/2015/11/2015-esma-1464_annex_i_-_draft_rts_and_its_on_mifid_ii_and_mifir.pdf

As seen in this announcement
http://www.thetradingmesh.com/pg/events/mike/read/848379/workshop-on-utc-traceability-for-the-global-financial-sector-realisation-of-a-unified-common-clock
the meetings on the subject, some of which have included staff from
BIPM, have been "no press or media will be in attendance".  Note the
admission in there that "At the microsecond level, the realisation of
a unified 'common clock' at the timestamp does not exist."

To the dismay of the BIPM the regulators have produced
https://www.esma.europa.eu/sites/default/files/library/2015-1909_guidelines_on_transaction_reporting_reference_data_order_record_keeping_and_clock_synchronisation.pdf
wherein it is allowed that GPS, GLONASS, or Galileo may be used for
the timestamps so long as there is a traceable connection.  That was
further expanded in
https://www.kaizenreporting.com/just-a-second/
which has since been made login-only, but which said
  On the back of the FCA's MiFID II conference this week we were with
  the National Physical Laboratory (NPL) to find out.  Their workshop
  "UTC traceability for the global financial sector, realisation of a
  unified common clock" attempted to answer some of these questions.
  Taking the front seat was MiFID II's RTS 25, which outlines the draft
  regulatory technical standards on clock synchronisation.  The direct
  impact of this piece of regulation was apparent from the range of
  professionals from all aspects of the industry including financial
  firms, exchanges, consultants and scientific organisations.  Most
  notable was the presence of ESMA and the Bureau International des
  Poids et Mesures (BIPM) as both have a heavy hand to play in RTS 25.
  We suggest firms take a look at Table 2 in the Annex of RTS 25 which
  sets out the level of granularity applicable to their trading
  activities.  It is important that firms are clear on what is expected
  of them and should be in advanced discussions around the technology to
  be adopted to meet the new regulatory obligations.

In some of the salepitches it has been evident that the answer to the
MiFID II and Warner Losh's question is that companies exist who will
install high-precision timestamping gear at the market and lease a
direct connection to the nearest national metrology institute.  Those
salespitches also indicate that despite the use of term UTC in the
text of the regulation, the underlying system will have to work in TAI
(which I take that they are simplifying the sales pitch and they
really mean a TAI-like timescale).

That way the initial trade timestamps will be a provisional local time.
When the regulators come knocking on the door asking about the actual
UTC of a transaction the provisional local time can be converted to
a value of UTC consistent with BIPM.

And all that is the same technique that astronomers and navigators
have used for centuries.  The new part in all the above leaked
information is that the way that timekeeping works has now been
inculcated to financial market regulators.

--
Steve Allen                    <sla at ucolick.org>              WGS-84 (GPS)
UCO/Lick Observatory--ISB 260  Natural Sciences II, Room 165  Lat  +36.99855
1156 High Street               Voice: +1 831 459 3046         Lng -122.06015
Santa Cruz, CA 95064           http://www.ucolick.org/~sla/   Hgt +250 m