In Michael Lewis’s 2013 best-seller Flash Boys, in which he takes a deep dive into the world of high-speed trading, he reveals how certain financial insiders had managed to game the system by reducing latency as much as possible in order to get an advantage over their competitors. Lewis showed how a fiberoptic connection that is just a millisecond or two faster than someone else’s can allow automated trading software to conduct transactions at the lowest possible prices. For the user (or machine) that had just the tiniest increase in latency, a stock price would appear on screen, but by the time the command to purchase it was received, the offering would no longer be available.
While high-speed trading represents an extreme case, and that level of optimization isn’t quite as big of a deal for the vast majority of market transactions, it underscores the challenge faced by all DevOps professionals in the financial services industry. Consumers across industries already have very little tolerance for poor performance, but their patience becomes even thinner when opportunities for financial gain are at stake. A few milliseconds usually won’t make much of a difference, but once the performance threshold is breached and customers’ portfolios are being negatively impacted, the backlash is going to be severe.
Therefore, it goes without saying that financial service providers have to take performance just as seriously – if not more so – as their counterparts in the ecommerce or media industries. This includes everyday omnichannel banking services such as transferring funds or checking balances but is even more imperative for online trading vendors, who must also factor in the ever-changing nature of the worldwide markets.
Such is the case for GAIN Capital, a financial services firm which runs a public trading platform on web infrastructure around the world that allows users to buy and sell stocks across different international markets. Because GAIN Capital operates as a SaaS provider connecting consumers to the financial markets, they cannot afford to be a performance bottleneck. And due to their emphasis on markets where digital experience monitoring can be difficult, such as China and Singapore, GAIN has an even greater challenge to deliver an optimized digital experience to their customers.
The complex nature of GAIN’s trading platform requires streaming HTTP service and their rest API to deliver real-time stock prices and updates, which means that they must run constant synthetic tests of those components through complex scripting scenarios, as well as detect and resolve any problems that arise in a matter of minutes.
GAIN Capital must also be able to collect this data from remote geographies in order to best emulate the kind of user experience that an end user would have when trying to conduct business on the financial markets in China, Singapore, the U.S., or anywhere else around the world. That means having optimized connections to both the market data, as well as the end users who are accessing GAIN’s trading platform and being able to detect performance issues on either their own internal servers or on the infrastructure components within their architecture, such as DNS and CDN providers in those given regions. After all, consumers don’t care if a problem lies with one of GAIN Capital’s servers or with a third-party vendor; they just want to be able to conduct their transactions in a timely manner.
At the end of the day, financial service providers have a lot on their plate considering that they must provide the best possible digital experiences across many different locations, while also dealing with the limits imposed on them by the requisite security measures that must accompany any connections in which consumers’ private and sensitive data is being transmitted. Failure to do so will impact not only their own revenue streams but those of their customers’ as well – and by extension, their own brand and reputation.