Time could be up for leap seconds later this year, if Tuesday night's addition of one second to the world's time proves to be a success.
International weights and measures standards-keepers at the Bureau of Weights and Measures (BIPM) near Paris want an end to leap seconds and favour a single, consistent time standard. BIPM has expressed concern about the risks associated with the leap second.
For example, there are four different satellite networks with their own separate time offsets, which the BIPM fears might cause confusion and possible disaster.
Elisa Felicitas Arias, director of the BIPM time department, told The Reg: "Tomorrow's leap second will be a good 'test-bed' in preparation for the final discussion."
Experts will meet at the World Radiocommunication Conference in November to discuss ending the practice of coupling of Universal Co-ordinated Time (UTC) with astronomical time – a practice that first gave us the concept of leap seconds.
The last confab debating the death of the leap second took place in January 2012 – and broke up when members failed to reach a decision.
Other measures being proposed for November's planned conference include having two reference time scales: one with leap seconds and the other without, thereby being continuous. This last option is against the meteorological practice, where the reference must be unique.
Slicing and dicing the time
The news came as cloud firms, money markets and space projects prepared to check their timepieces on Tuesday night and Wednesday morning when the 25th leap second will be added to UTC. Those who have built their trading and galactic empires on Linux will be most anxious – but with little justification.
When the last leap second was introduced in June 2012, the Tux reservation systems of travel giant Amadeus choked on a long-standing kernel data bug. That bug had hidden in Linux kernels version numbers 2.2.26 to 3.3, inclusive.
Problems were reported to have affected Mozilla, StumbleUpon, Yelp, FourSquare, Reddit and LinkedIn, too. Red Hat – at least – released a patch. Mozilla’s Hadoop and ElasticSearch servers were also taken down by a date-sensitive fault in that system, too. A Java fix was quickly developed.
Thunderstorms ahead for the cloud world
Amazon and Google didn't report problems in 2012, but three years on they have put in place elaborate mathematical, and temporal fixes to avoid potential disruptions. Both firms will implement a policy of cutting up the second and feeding it into the clocks on tens of thousands of their servers 12 hours either side of midnight.
That means both AWS and Google’s clouds will implement non-standard time and exist – briefly – in their own time zones.
Microsoft is going for the big bang approach. Azure servers will adjust to the new second based on the time zone they’re in.
Less clear is how those running the world’s networks of other PCs – such as cloud service providers and financial markets – will cope.
The only ones not taking any chances will be international space projects: leap seconds are introduced on either 30 June or 31 December, dates when space missions deliberately do not launch just in case of problems.
The second in UTC is defined as the length of time taken for 9,192,631,770 cycles of the caesium 133 atom produced by the transition between two levels.
While UTC might be perfect in that it’s accurate for a period of 158 million years, humans prefer to measure time using astronomical time. This is slowing down thanks to the slowing rotation of the Earth and is updated about every one to three years.
The leap second keeps UTC in step with astronomical time. ®