The average high-frequency trading algorithm can process trade orders in under 400 microseconds, which is 1,000 times faster than the blink of an eye. Survival in an industry where every fraction of a second matters means high-frequency trading (HFT) firms must exploit even the slightest speed advantages to gain an edge over the competition.
Minimizing system latency has become the primary way to achieve these slight accelerations in transaction speed, causing high-frequency traders to spend heavily on the fastest computing technologies and data center real estate in close proximity to exchanges. This co-location of an HFT firm’s supercomputers in the same data center as exchange servers helps traders achieve the fastest possible processing speeds and quickly use data-driven decision-making to turn a profit.
Positioning their equipment just a few feet of cable away from exchange servers enables HFT firms to capture the most up-to-date market pricing a few milliseconds before the rest of the investing community. This real-time insight into market dynamics offers traders a distinct advantage as they are able to analyze and react to the current state of the market long – actually, a few of milliseconds – before everyone else.
Co-location has become a lucrative proposition for exchanges, who charge HFT firms exorbitant rental rates for low-latency positions in their data centers. This insatiable demand for data center space is the primary reason why many stock exchanges are rapidly expanding the size of their data centers. For example, while the old New York Stock Exchange building was only 46,000 square feet, the new data center in New Jersey sprawls over 400,000 square feet and features 60,000 square feet of co-location space for low-latency trading opportunities by hedge funds and investment firms.
If a millisecond of trade delay seems insignificant, consider this example from the South African Financial Markets Journal: A trader in Sandton, a town just down the road from the Johannesburg Stock Exchange (JSE), could receive packets of updated trading pricing and volume data in as little as 2.7 milliseconds. Conversely, a trader located in Cape Town, almost 900 miles away from the JSE, wouldn’t receive that same packet of information for approximately 20 milliseconds. Simply because of geographical proximity to the exchange platform, the trader in Sandton will have a 17.3 millisecond advantage over the trader in Cape Town.
In the high-speed world of HFT, this advantage could render the Cape Town trader’s information extremely outdated and potentially useless.
There are challenges to co-location particularly for the smaller HFT houses. Cost is a chief obstacle, as exchanges continually increase rates for co-location space, particularly in big cities close to exchanges where data center space is at a premium. Sourcing, deploying, and maintaining the data infrastructures involved with co-location are sometimes outside the budget and expertise of smaller HFT firms, as is the complexity of managing the availability and security of such business-critical equipment off-site.
However, hardware solutions are available now which are optimized for easy co-location data center deployments. Server configurations that are right-sized to fit the standard power capacities of most exchange co-location facilities and have a small footprint to economize on space help reduce the costs and complexity of co-location deployments, while ample processing power adds to the fight against system latency.
Rapidly advancing computing and connectivity technologies are permanently altering the trading landscape – from open outcry to quiet automation, from men to machines. In this environment of ever-increasing trading speeds, HFT firms must find ways to compete more intelligently while also reducing the costs and complexity of their IT infrastructure. Co-location is an essential component of infrastructure strategy as firms seek to exploit microsecond latency advantages to help them succeed in today’s increasingly competitive financial markets.
To learn more about how co-location can improve your business, I invite you to follow me on Twitter at @sureshaswani2.