By: Jason England, Head of Capital Markets Networks and Hosting, TD Securities
Imagine if you could get prices and data 10 milliseconds faster than your competitors. The arbitrage benefits of the lowest latency transatlantic link would be enormous. NYC to London in 22 milliseconds is possible via radio and is a full 50% faster than the fastest cable available today.
Electronic trading has automated and increased the efficiency of markets across multiple asset classes and regions. High-performance computing has enabled trading to occur in microseconds. Faster connectivity has created arbitrage opportunities and such is the ability for systems to exploit minor latency variations that exchanges often ensure all participants are equidistant using bundles of fibre.
In the past 10 years, markets have developed to optimise connectivity. MPLS and exchange switched connections have been abandoned and replaced with fibre for faster connections. Network paths have been dug in straight lines to optimise paths. More recently, we have seen the adoption of microwave and lasers to reduce connectivity times.
The so-called ‘Race to Zero’ aims to eliminate any delay over and above that which can be achieved through the laws of physics. Quantum techniques may yet yield true light speed communication at a distance. However, we are many years from commercial offerings.
The financial markets of London and New York are currently separated by a mere 33 milliseconds via the current lowest latency transatlantic link. That is equivalent to the duration of one and a half blinks of an eye. That time is achieved through fibre optics linking trading venues in New York and in London at the somewhat sedate two-thirds of the speed of light. There is some ability to optimise this through wavelength selection, but the improvements are marginal. Improvements to the overall duration have been attempted through the deployment of microwave and laser systems to achieve near light speed on the land portion of the links. There are, however, significant portions of the transatlantic fibre link that exist under the ocean and cannot be further optimised with current technologies.
Satellites are not the answer, well at least not current geostationary ones. They sit so far out in space that they create a latency of many hundreds of milliseconds. Lower orbit satellites remain impractical due to fueling requirements, however, high altitude balloons or drones are potentially worth investigation.
Speed of light transatlantic communications are not new, in fact, they have been achieved regularly and reliably for over 100 years. Marconi first achieved Morse code communication using massive spark gap transmissions that could be detected thousands of miles away. In the years since Marconi, we have been pursuing bandwidth almost exclusively. The explosion of data in the Information Age has driven the deployment of undersea cables as a cost efficient and reliable mechanism for the exchange of data. Some enhancements have been made, but fundamentally the major portion of the link is slower than when Marconi tapped out Morse code.
Current microwave links rely on line of sight communication. Light and microwave radio waves don’t bend and thus can’t reach targets beyond the visible horizon. Generally, these systems operate in the region of a few miles to perhaps 100 miles in ideal conditions.
There are, however, radio frequencies that are able to propagate beyond the horizon. Shortwave radio transmissions are able to reach over the horizon through a very similar mechanism to light in a fibre.
Shortwave radio frequencies bounce between the transmitter source and the destination, reflecting off the ionosphere and the ground. These radio waves also travel at near light speed, crossing the Atlantic in a little over 20 milliseconds. Renewed research in radio communications for over the horizon communications spiked once the Chinese military shot down a satellite on Jan 11, 2007. This event highlighted a weakness in the satellite as a maneuverable WAN device. Defense contractors are investing significant sums in developing platforms to provide operating theatres with reliable data communications at distances of many thousands of miles.
Defense contractors and civilian providers offer systems today that provide data communication over ranges of hundreds of miles. Increasingly, commercially available incredibly sensitive software defined radio appliances are creating the potential for a transformational event in transatlantic trading event latency.
So why isn’t everyone using this?
Attributes of shortwave and radio transmission in general restrict the volume of data and thus reduce the number of use cases for which it could be used.
Performance and latency tuners tend to focus on improving what they can directly influence, the pursuit of incremental improvement and its value is well articulated by my colleague David Sewell in his blog about marginal gains in IT.
Occasionally we see step change developments which appear obvious in hindsight. The lowest latency transatlantic link exists today, it’s existed since Marconi and just needs some refinement to transform the arbitrage markets.