If asked to name the differentiators for global investment banks trading into and across Asia, you might be forgiven for not having network optimisation at the top of that list. Yet, in reality, the entire end user experience depends on the underlying performance of the network. Many banks and brokers struggle to ever achieve the same quality of execution service in Asia that they deliver to clients elsewhere, leaving the few who do to dominate the region's fast growing market for cross asset execution.
This isn't good for the region or the market's participants. Asia is already a very challenging region in which to operate, with softening volumes and high volatility making recent conditions especially difficult. Anything that enhances liquidity or improves execution quality will find an eager market.
Compared to trading in Europe or the U.S., the sheer logistics of trading into the Asian markets make it a hugely expensive proposition, largely due to its diverse geography. Unlike the U.S. or even Europe, it's a fragmented market, spread over a number of island nations with thousands of miles in between. To get from Europe, trades need to travel a wire that takes in Russia, the Middle East, or Africa. This not only adds to the distance travelled but also introduces geopolitical risk. These are opaque countries; opacity creates risk and risk creates cost.
As such, participation has historically been limited to either those with sufficient scale to shoulder the required investment on a global basis, or the intra-regional participants (who still have to deal with the fragmented nature of the market, even if not the transcontinental links) for whom it's crucial to be able to demonstrate participation in all parts of the region.
For trading companies trying to weather looming economic storm clouds, such as fading Chinese stimulus and weakening global demand, overall profitability in the region is increasingly difficult to sustain. In some cases, businesses are questioning decisions to introduce new products or data feeds based on the cost of the required technical infrastructure.
The network effect
So why is this such an issue and how is the network a possible answer? Surely the cost of cable is the cost of cable: a necessary business expense?
In some ways, yes. The fragmented nature of the Asian market and the long distances between hubs is unavoidable and exposes (relatively) fixed costs when it comes to connectivity. However, given the history of poor quality services in this region, there is a definite and significant market opportunity for anybody able to differentiate themselves with something better.
So perhaps it's worth considering whether network costs are really as fixed as they seem. If a firm can both deliver optimised performance and free up bandwidth to invest in new services, it will be in a winning position. An everyday technical challenge could suddenly be a key differentiator for those operating in the region.
The first step to doing so is to invest in a level of close scrutiny and control of bandwidth usage more stringent than firms are accustomed to. Frequent, millisecond sampling rates can capture activity bursts that would have gone unnoticed between samples in less granular traditional architectures. If unnoticed, the bank might then believe that it's sitting pretty with comfortable 10-15% utilisation when in fact it's experiencing a burst in excess of 70-80%, leading to slowness and lost data packets for mission-critical trading applications.
Forensic, holistic detail can identify which services are causing the bursts - often to find they're not critical and shouldn't be on the expensive, low-latency network at all.
Re-routing non critical traffic to other, less expensive networks and ensuring events such as batch transfers or daily updates take place outside of the trading window means banks can optimise the performance of their lowest-latency, most expensive networks. This also frees up bandwidth for new services to attract more clients and help gain market share.
Of course, optimised network bandwidth is good practice anywhere, but it's an especial boon to trading across Asia. Large scale executing brokers often use expensive, high-bandwidth lines between their main hubs (Hong Kong, Tokyo, Singapore) and London and New York.
While it's acknowledged that relative bandwidth cost goes down for higher capacity investments, adding extra capacity rather than optimising the existing network is still expensive.
For example, 300Mb lines can cost in the region of £10,000 (U.S. $12,971) per month. Upgrading to 500Mb can instantly increase that cost by as much as £60,000 (U.S. $77,826) per year. That's an unpalatable prospect given the fast growing market data bandwidth requirements, increased regulatory accountability, and soaring operational-level costs which collectively continue to increase the service level burden to banks.
Finally, it's worth noting this isn't just a problem for the investment banks and brokers. The lack of new products (and cost pressures on existing ones) continues to suffocate new market development. As Asian capital markets look to continue their march of progress and stand shoulder to shoulder with the U.S. and Europe, better network instrumentation and optimisation will be vital to surmounting the region's unique barriers.