A common theme I hear from many senior IT people who run FX trading platforms is how the priority of rejuvenating their FX infrastructure is increasing. Why now? All of the big investment banks and FX ECNs have operated electronic FX trading platforms for some time now — since back when much of the trading demand from their clients wasn’t electronic, which meant latency requirements weren’t that stringent and peak volumes weren’t as high as they are today. There were also fewer sources of liquidity, less sophisticated trading capabilities, a lot less regulation to worry about, and corporate risk officers pretty much left them alone.

Many of these platforms used (and still use) several different types of messaging systems to do what they needed: multicast for price distribution, software brokers for persistence, and web streaming software if they wanted to offer a single dealer platform (SDP). So the infrastructure started off complex, which doesn’t help application architectures stay simple, and we all know that over time architectures grow more complex until there is a dedicated effort to simplify.

Fast forward to today and some firms who haven’t invested in their infrastructure for years are finding that their clients aren’t just fully electronic, but have leapfrogged them to such a degree that traffic bursts are overwhelming their systems and they can’t offer the low, consistent latency customers expect. The fact that their inherently complex platform has grown over time to accommodate features and capacity, along with increasing demands from the regulators, internal risk officers and security auditors, means their application architecture is VERY different from where it started so long ago — frequently approaching the infamous Big Ball of Mud.

The focus on features, algorithms, regulatory compliance and risk management has sometimes led to situations where the infrastructure, including messaging, has been neglected. So firms now find themselves in a situation where their old messaging infrastructure is causing a variety of business problems that they can no longer ignore, such as:

  • Poor — or at least unpredictable — system performance, especially during peaks.
  • System fragility causing too many outages due to infrastructure complexity, components not handling overload well, or the use of multicast in places where it should never go such as to desktops and over the WAN.
  • Outages that take too long to troubleshoot due to architectural complexity and lack of proper management visibility.
  • New feature development is too expensive and hard to predict because of the workarounds required due to complexity and shortcomings in the infrastructure.
  • It’s too difficult and expensive to keep the system running. Too many servers and software licenses, and it takes too many specialized experts to maintain, tune, troubleshoot the infrastructure and its complex interactions with the applications.
  • Lack of governance/production inventory/map to even know what applications are producing or consuming what information, at what rates, where they are deployed, what hardware or software they are using.
  • It’s too difficult to pass internal security audits (especially when using multicast) and you can only get exceptions for so long.

We have many clients who have or are in the process of replacing literally hundreds of message brokers with a much smaller number of our appliances to eradicate the complexity of multicast and the more general problem of too many moving pieces from their FX trading platforms. They see faster and more predictable performance with headroom to scale, higher availability, more robust system governance, better WAN performance and utilization, better management and troubleshooting, elimination of cascading failures, passing security audits and the luxury of being able to focus on their business applications without constantly building workarounds or dealing with production problems.

If your FX trading infrastructure is getting a bit long in the tooth or you just want to see the type of simplified architecture that other leading banks have adopted, read our new FX whitepaper and please tell us what you think.

Shawn McAllister
Shawn McAllister
Chief Technology Officer & Chief Product Officer

Shawn McAllister is responsible for the strategy and delivery of the Solace PubSub+ Event Streaming and Management Platform. He leads a team of incredibly talented engineers and architects in this endeavor Shawn has worked with many of our clients to help them adopt event-driven architecture and to learn first-hand their needs as input to the innovation built into the PubSub+ Platform. He has participated in the definition of various OASIS messaging protocol standards, including MQTT 3.1.1, MQTT 5.0, and AMQP1.0.

Before joining Solace, Shawn led software, hardware, and test engineering teams at Newbridge Networks (later Alcatel Canada), where he was responsible for developing features on ATM and Ethernet switches as well as the 7750 Multiservice IP Router. Shawn holds a Bachelor of Mathematics from the University of Waterloo, with majors in both Computer Science and Combinatorics/Optimization.