James Sinclair CEO MarketFactory,Inc.
James Sinclair CEO MarketFactory,Inc.

Distance Latency: Overcoming the constraints of Geography in FX

Two trends conspire to challenge established technologies used by banks, hedge funds and any institution trading actively in financial markets, including foreign exchange. The first is the increasing speed with which computers make algorithmic trading decisions, quote and trade; the second is globalization meaning that traders must transact on multiple ECNs, exchanges and other venues. These ECNs are dispersed across major trading centers including London, New York, Chicago and Tokyo. ‘Distance latency’, the time taken for a message to travel between trading centers, is an increasing issue for traders but is governed by an immutable constant – the speed of light.

First Published: e-Forex Magazine 33 / Algorithmic FX Trading / October, 2008

The market price may change in the time it takes for an order to travel from London to New York or even from Chicago to New York. Prices on ECNs in one center may be out of date by the time traders in another center see them, let alone attempt to transact on them. Distance Matters Conventional approaches focus on collocating trading models with an exchange, embedding software in hardware, high-performance computing and other means that are certainly valuable, if not essential. However, this does not solve the problem for traders whose models need to deal on multiple venues. Distance latency dwarfs all other forms of latency. Figure 1 Figure 1 shows the comparative latencies involved in trading on venues in various centers. The blue bar on the left, barely visible, is the trading model latency, assumed to be 1 ms. The yellow bar is the ECN turnaround time, assumed to be 10ms though it varies greatly between ECNs and in different circumstances. By far the longest form of latency is the...continued

Exclusive Content

The full article is only available to current subscribers. Click here to sign in or subscribe by clicking here