Einstein and intergalactic time synchronisation

Latest News AND EVENTS

Stay up to date and find all the latest news and latest events from Metamako right here.

Einstein and intergalactic time synchronisation

Posted on September 07, 2016 by Ciaran Kennedy 07 September 2016

David Snowdon, Founder and CTO of the ultra low latency trading technology firm Metamako, takes on Albert Einstein's theories of relativity.

Timesyncing and Timestamping for low latency trading intergalactic

A while ago I noticed this article suggesting that precise timing and timestamping in modern financial markets is a futile endeavour, since Einstein's theories imply that the ordering of events is impossible to determine. 

Is this really true!? Turning for advice to a trusted cosmologist (turned big data expert), he informed me that this is actually the current understanding in modern physics. Imagine my dismay, having built successful clock synchronisation products to be used for monitoring and logging the events occurring in markets, finally putting an end to all the discussion of front running and unfair advantages. It didn't make sense to me. We'd seen how effectively synchronised clocks can can provide precise low latency time stamping for trading operations.

How did Einstein think about time synchronisation?

Of course, the author of that article also pointed out the effects that the uncertainty in the ordering of events was really a function of the relativistic effects of time dilation. But it turns out that the relative movement of, say, London and New York is pretty darn close to zero – say a few centimetres per year. Gravity would also have an effect on time, but in the words of my trusted cosmologist, "If London was built on a nugget of Neutron star" ... you'd have something to consider.

A neutron star is the collapsed core of a massive star, and is one of the densest objects in the universe thereby creating intense gravity. 

That means that since all of the observers of our markets have the same frame of reference – i.e. they are not moving (at significant speeds) relative to each other and experience the same gravitational field – they can each build a picture of the ordering of events. 

For example, if an event occurs in London, which is roughly 5,600 km from New York, then an observer in New York would see it approximately 18.6 ms later in free space. It turns out that light travels slower than that in a fibre, and fibres generally don't run in straight lines. But these are things which can be measured – and in fact, that's exactly what time synchronisation protocols like PTP do – send a ping, wait for a pong, time the difference and divide by two. 

That's the basis of how organisations like NPL determine International Atomic Time – a reference time generated by synchronising of hundreds of clocks across the world – which is used to determine UTC.

How does this apply to low-latency trading and regulatory compliance?

Clocks can easily be synchronised to nanosecond-or-better precision, and this is routinely used in applications like air traffic control, GPS, and scientific experiments like those at CERN. As a real-world example, Metamako has shown that it can achieve very precise synchronisation with MetaWatch.

So, did Einstein miss the boat on Timestamping?

Einstein was right, but the impact on our day-to-day lives is so small that it's negligible. Just as you don't care about the aging effects of air travel (the lowered gravity at a cruising altitude results in a slightly increased rate of time for airline passengers), we don't need to consider the effects of the relative movement of London and New York in our time synchronisation algorithms. 

Even if we were, say, orbiting the earth – we'd still be able to understand, measure and calibrate out the effects of the change in gravity and relative speed. GPS wouldn't work without it! GPS relies on time synchronisation between the satellites, and a GPS survey accuracy of centimetres or millimetres is routine... a distance travelled by light in substantially less than a nanosecond (one billionth of a second).

If we ever get to the point where we're on interstellar space ships where we are not able to determine our relative velocity, or gravity, and if we wish to trade and observe others trading, we'd really have to start to considering the idea of relativistic simultaneity. 

So spare a thought for the future generations of intergalactic traders who will need to deal with that mess. We're fortunate enough to have an excellent way to measure the timing and order of events in our earthly markets.