While latency figures between market centers in the US and Europe have tumbled over the past decade as exchanges and co-location and network providers invested in high-speed, high-bandwidth connectivity to support the demands of algorithmic and high-frequency traders, latency across the Asia-Pacific region has generally remained high, creating an impediment to adoption of high-speed trading strategies in the region.

This has created inefficiencies between disconnected markets that traders would be keen to exploit, if they were able to gain a latency advantage—even a fairly small one, by Western standards.

“In Europe and North America, the market data ecosystem, particularly provided by the stock exchanges, is far more mature than it is in Asia. So generally, you haven’t seen the HFT fund environment within the region be as robust or mature because the data hasn’t been there to support those trading strategies. But that is changing,” says Benjamin Quinlan, CEO and managing partner at Hong Kong-based strategic consultancy firm Quinlan and Associates.

Indeed, the data is available. The challenge is getting it from a market in one regional center to a trader in another fast enough to make trading attractive.

Part of the challenge is the sheer scale of the region. While traders in London argue over connectivity between datacenters a few miles apart around the greater London area, or while firms in New York struggle to shave fractions of milliseconds off the latency between facilities in New Jersey and those in Chicago, just under 800 miles away, firms in Asia have to contend with much larger distances between markets—often without the option of over-ground fiber routes or even-faster wireless microwave data transmission, instead having to rely substantially on undersea fiber between market centers in different countries.

For example, the distance from the Japan Exchange Group, operator of the Tokyo and Osaka stock exchanges, to the Sydney-based Australian Securities Exchange (ASX) is more than 4,800 miles. The distance from Tokyo to Singapore, home of the Singapore Exchange (SGX) is about 3,300 miles. And the distance between SGX and ASX is almost 4,000 miles. Achieving the kind of meaningful latency reductions over these distances that would incentivize more high-frequency trading is a huge challenge for market participants.

Arthur Rank, global director for capital markets solutions at network provider and market data carrier Colt Technology Services, says he believes the vendor’s latest network update between SGX and ASX is meaningful enough to benefit HFTs trading derivatives and foreign exchange. The vendor will next week announce that it has optimized its SGX–ASX route to below 87 milliseconds roundtrip latency between the market centers for HFT clients that it hosts in both locations.

However, because of the way networks in the region are constructed, this latency improvement also underpins the co-location offerings Colt has across the Asia-Pacific region. “If you want to get data from Australia to Hong Kong, or Tokyo, the fastest route will always be via Singapore. So, this is part of that bigger ecosystem upgrade. If a customer in Hong Kong wants to access grain derivatives data from ASX, then improving the route between SGX will also improve the delivery lead times on to Hong Kong and Tokyo potentially,” Rank says.

The latest reduction is part of long-running efforts to improve latency across the region, including recent upgrades and additions to Colt’s other ultra-low latency routes, such as new ultra-low latency routes between the Stock Exchange of Thailand, Hong Kong Stock Exchange (HKEX) and SGX, Japan Exchange Group (JPX) and HKEX, HKEX to SGX, and ASX to CME. It also added connectivity in South Korea and Taiwan. In addition, in 2020, Colt implemented the Arista 7130 Layer 1 switches (previously known as Metamako Metamux) within HKEX, allowing the lowest possible latency via market switching without needing to invest in or manage Layer 1 switches themselves.

Beyond the size of the region, another impediment to greater adoption of cross-border automated trading is that regional exchanges may be less focused on developing products for the nascent HFT market, which their Western counterparts are already exploiting to deliver revenue increases. For example, New York-based Nasdaq reported Q2 revenues of $283 million for its Investment Intelligence segment, which includes its market data, index, and analytics business, representing about 32% of total exchange revenues. Comparatively, SGX reported S$70.7 million ($51.4 million) for its data, connectivity and indices business during Q2—a lower number reflecting its comparative size, but also a lower percentage: At SGX, data-related revenues were just 14% of total exchange revenue.

“Most of the revenues from Asian exchanges still come from listing and trading, as opposed to the sale of market data (such as real-time data and indices) that you’ve seen from Nasdaq, LSEG and NYSE. The Western exchanges are far more advanced along their data journey. And with data availability comes the ability to do more automated trading and high-frequency trading,” says Quinlan.

As that ability becomes more viable, the focus on intra-region latency will increase, forcing exchanges and network providers to up their game further. In Colt’s case, to manage its ongoing latency-reduction efforts, the vendor has a global ultra low-latency network and a product team continuously looking at optimizing routes and improvements it can make to its own backbone, as well as in partnerships with other providers to “straighten out lines from a technology perspective” and in the switching space to improve latency and throughput.

“It’s an ever-evolving game where we’re continuing to lay new fiber routes and we also work in partnership with a number of other providers to be able to offer that best-in-class offering in ultra low-latency,” Rank says.