The International Earth Rotation and Reference Systems Service, a French agency, has announced a leap second at the end of June 2015 – and if you don’t know what that means, you’re not alone.
At the root of the issue is one simple fact: the Earth’s rotation is slowing down. Many things, including geological and climactic changes, can affect our planet's spin, but the most salient one is the moon, whose gravity produces a drag on our planet, meaning that, if all else is equal, each "solar day" will be ever-so-slightly longer than the one before it. As the solar day slows, just about every timepiece in the world, except for the sundials, will fall out of sync.
To compensate for this, the French can impose an extra second – a leap second – to be added at the end of any month. It occurs just before midnight – clocks read an otherwise nonsensical 23:59:60, and then everything goes back to normal.
And in most cases, it’s not a big deal. In fact, the leap second went largely unnoticed by the general public when it was introduced in 1972. But with each subsequent leap second – there have been a total of 25 since then – society's reliance on computers has increased.
Many computers use Network Time Protocol, a service that periodically checks the computer’s internal clock against against an external, more accurate one. But when NPT presents a computer with a time that theoretically shouldn’t exist, the computer can become "confused," and crash.
That's what happened on June 31, 2012, when we last had a leap second. That night, a slew of major websites – including Reddit, Gawker, and Yelp – experienced outages caused by the leap second. Worse, some 400 Qantas Airlines flights were delayed for two hours because the leap second bug had shut down their reservation system.
For other organizations, the leap second effect was not so dire. Google avoided meltdown that year with what they called a “leap smear.” Instead of introducing a redundant second all at once, Google modified their NPT servers to add the extra time in millisecond increments over the course of a day. By the time the actual leap second occurred, Google's servers had already taken it into account, thereby preventing a crash. The workaround was actually first tested in 2008, and the tech giant has been using it to account for leap seconds ever since.
In a 2011 blog post, Google engineer Christopher Pascoe called the method, “one of our coolest workarounds, that took a lot of experimentation and verification, but paid off by ultimately saving us massive amounts of time and energy in inspecting and refactoring code. It meant that we didn’t have to sweep our entire (large) codebase, and Google engineers developing code don’t have to worry about leap seconds.”
Leap second outages have prompted a very different reaction in other technological communities: calls for the outright removal of the leap second. This wouldn’t be immediately problematic – the time difference would be minuscule in the short term – but bigger issues could come up in the distant future. Without the leap second to keep machine time and solar time in sync, in tens of thousands of years, noon would fall at night and New England would get snowy summers. Alternatively, computers could switch to a different time standard, but the coding necessary to make that change might be more difficult than developing a workaround in the first place.
As it stands, there’s not much else that can be done to solve the leap second conundrum, as syncing up the rotational and conventional time standards would literally require moving the Earth.