FX looks to exploit the scalability benefits of cloud computing

This content was originally published by e-Forex.net.

Institutional FX has always stood ready to adopt innovative technology. However, the complex nature of the market and its instruments have usually posed significant challenges for firms to overcome. The cloud is no exception. Vivek Shankar investigates.

A survey by McKinsey in 2022 revealed that just 13% of financial services firms had moved more than half of their IT footprints to the cloud. While this survey wasn’t specific to FX, it’s not hard to imagine a similar pattern playing out.

Some challenges like the prevalence of voice trading and security concerns come to mind. However, cloud technology has progressed to the point where not adopting it will likely put firms behind the curve.

“At its core, cloud computing delivers a more flexible and scalable computing resource capability, particularly if accessed through an on-demand, pay-as-you-go, ‘as a service’ model,” says Tim Carmody, Chief Technology Officer, IPC Systems. “Unquestionably, the flexibility and elasticity of cloud-based computing deliver benefits in terms of scaling, reduced CapEx, and faster go-to-market times for many workflows.”

So how has cloud computing evolved recently and how does it tie into most firms’ FX needs?

Cloud adoption holds several advantages for FX

“Financial institutions are having to cope with more regulations, limited resources, rising complexity, and increasing data volumes,” says Udai Abburi, Senior Director and Head of enterprise solutions and technology at ION Markets (FX). “Cloud computing offers a strategic solution for FX firms, enabling greater innovation, automation, and compliance.”

He explains that the cloud’s maturity as a technology makes it an ideal fit for several FX use cases. “As AWS calls it, the cloud takes away “undifferentiated heavy lifting”, and allows firms to focus on core business needs and value generation,” he says. “It democratises the playing field, lowers barriers to entry, and allows all firms to make use of sophisticated and scalable infrastructure across compute, storage, and networking.”

Vikas Srivastava, Chief Revenue Officer at Integral, says that the cloud offers significant workflow automation which gives rise to efficient processes and a reduction in errors and delays. “The ability to scale up and down based on specific needs at a certain time makes cloud computing an essential tool for modern FX trading operations,” he says. “For those who provide technology to the FX industry, making solutions available on the cloud allows firms to go to market quicker with new solutions, and for banks it is the fastest means of distributing FX liquidity electronically, either through a multi- or a single-dealer platform or via an API.”

Despite these advantages, Carmody concedes that the cloud may not present itself as a primary option for some FX trading firms. “FX firms might not necessarily want elastic computing on demand for core trading requirements, where a more deterministic and guaranteed strategy is crucial in times of market volatility,” he says.

However, firms might benefit from the cloud’s flexibility and elasticity in on-demand services that feed into trade execution. “FX may be the poster child for hybrid computing since the data used to inform trade execution is broad and deep, far beyond just financial markets and embracing political, economic, weather, and all sorts of other information from vast numbers of sources,” Carmody says. “Elastic computing and AI can certainly help to garner and manage all this data.”

Abburi echoes the point about elasticity and says cloud computing infrastructure readily allows for on-demand provisioning, elasticity, availability, scalability, reliability, security, compliance, and reduced total cost of ownership.

“Hybrid cloud options and cloud-agnostic platforms that can be incorporated into cloud-native applications are readily available,” he explains. “The cloud delivers fit-for-purpose compute capabilities, large-scale clusters, and low-latency network architecture.”

Srivastava points to scaling as a particular benefit—especially given its impact on costs.

“Scaling isn’t just about the number of transactions or throughput, scaling is also about growing a firm’s footprint across customer segments, geographies, and product offerings,” he says. “A trading firm can create a new product and easily expand infrastructure, and products can be differentiated to meet regional needs and easily expanded up or down on a case-by-case basis across geographies. Whereas, on-premise solutions require additional hardware, network, and staff to run operations.”

Despite architectural complexity, the advantages firms gain when speaking of scaling are enormous. Once again, Carmody cautions that the cloud won’t solve bad data issues despite all the advances the technology has made.

“It doesn’t matter how much data can be consumed and processed in the cloud if it’s bad data,” he says. This makes the choice of cloud services provider critical. Carmody points to IPC’s partnership with Celoxica as an example of a collaboration that adds value to firms.

“We collaborate with specialist data businesses like Celoxica to deliver ultra-low latency, normalised market data solutions to market participants, accessible to all users within our extensive global trading network,” he says. “Any third-party business has to pass our due diligence standards to qualify as an IPC partner or collaborator. It’s the old adage, we do this work so you don’t have to.”

Customizability is another advantage that stands out when speaking of the cloud—something Abburi stresses. “Cloud solutions allow for globally distributed infrastructure and related platform capabilities, not only providing global reach and distribution but also catering to multi-region high-availability (HA) and disaster recovery (DR),” he says.

“The infrastructure is on-demand, where you need it, when you need it,” he continues. “This allows for rapid innovation, experimentation, and fail-fast strategies.” The cost savings are also significant. Abburi points out that firms don’t have to build expensive data centres or hire expertise to manage complex infrastructure.

“The cloud provides a plethora of managed services, inclusive of making available open-source technologies across various types of databases, container orchestration, web and API gateways, and messaging services, providing key foundational building blocks upon which modern applications can be designed and deployed,” he says.

The cloud powers business growth

Cost savings and efficiency are one part of the ROI equation. Revenue and growth are the other. Carmody stresses that the cloud holds plenty of potential for firms to grow operations and serve new clients.

“As your trading strategies evolve, if you want to test a new market or spin up a new product capability, a hybrid model, with some element based on public cloud, is definitely a fast and cost-effective option,” he says. “You may want to change the mix of public cloud resources and more deterministic, guaranteed, and low latency delivery elements.”

He cautions against the use of a public cloud for competitive reasons. “It’s fair to say that your competitors will come after a trading strategy once they see you’re making money from it,” he says. “As such, elastic computing, on a public cloud, utilising shared resources, may not be the best model for maintaining a competitive edge.”

He cites the example of algo traders in this regard. “There’s a whole group who focus on reverse engineering trading strategies, or using AI and LLM to infer how they work,” Carmody says. “Being able to access and model data is one thing; inferring the thought process behind the strategy takes it to a whole new level. The minute too many other people can see what and how your algos are doing, your competitors also know what you’re doing.”

Competition is not the only shortcoming of public cloud services currently. Most clouds lack a critical mass of liquidity at low latency—a function of relatively few venues making themselves available on public cloud infrastructure. However, this trend is changing, and one expects this issue to resolve itself.

Even if low liquidity levels are a hurdle to executing trading strategies, other portions of the FX execution chain are not affected by this shortcoming. As such, nothing prevents firms from migrating those functions to the cloud.

Carmody believes scaling into the cloud is best carried out with a core trading strategy based on deterministic, guaranteed, ultra-low latency technology and connectivity. “Data from myriad sources already lives in the cloud, and elastic computing to inform trading decisions is great, but trade execution is still going to need to be on guaranteed hardware or via guaranteed ‘remote’ capabilities,” he says.

These cautions notwithstanding, Abburi points out that expansion is built into the cloud, thanks to service providers obtaining all necessary licences and compliance norms from local authorities when launching a service. These moves make scaling seamless, helping firms quickly deploy and test strategies.

He points to AI and ML advances as an example of how firms can quickly test viability and move forward. “Public cloud services allow for rapid experimentation and fail-fast strategies, without large upfront cap-ex costs,” Abburi says.

When put together, the benefits for firms span from risk management and post-trade processing to trade surveillance and fraud detection. In essence, the cloud helps firms advance their digital transformation programs, turning data into a critical asset that firms can leverage.

“Cloud technology serves as a levelling force for firms operating in FX and capital markets globally,” Srivastava says. “The reduced costs and access to high-quality tech from trusted partners empower businesses to remain competitive in an increasingly difficult market environment.”

Abburi says the cloud can simplify new trading desk priorities and offers an example. “There has been increased emphasis on liquidity aggregation and the need for real-time insight into various trading metrics on the trading desk,” he says. “These metrics include volumes across venues, fill ratios, price variations, and determining market impact.”

With applications deployed on the cloud, firms can quickly access real-time dashboards that visually represent data across customer channels and execution venues. When asked about the different data architecture options the cloud offers, Abburi reels off a list of features.

“A wide range of SQL and NoSQL databases, including time series databases, various configurable tools to create and maintain a unified data architecture, efficient compute and storage capabilities that can efficiently be managed and scaled, sophisticated tools such as search and analytics capabilities for processing large volumes of data across both batch and real-time solutions, etc.”

This list doesn’t include interoperability options that get different systems talking to each other and the ability to deploy the latest technology quickly.

Compliance and disaster recovery

Cybersecurity has become a major point of regulatory focus over the past few years. With incident disclosures now being treated as material information, financial firms face more scrutiny over their technological infrastructure than ever.

This focus has led to the rise of security certifications and standards like ISO 27001, ISO 22301, SOC 1, 2, and 3, NIST 800-171, GDPR, PCI-DSS, and Reg SCI. These certifications effectively act as a trust label for firms when evaluating a cloud services partner.

Abburi says that the scale of the average cloud services provider’s infrastructure delivers a significant benefit that individual firms cannot match. “Investments made by the cloud providers and economies of scale there in realised are far more extensive than what any one institution can achieve independently,” he says. “Strong authentication, authorization, secure data-in-motion and at rest, web security, network security, and monitoring are provided as built-in services.”

Tier 1 banks are not the only entities adopting the cloud as trust in security certifications grows. There are also specific NYSE and NASDAQ solutions that operate almost entirely in the cloud while adhering to the most stringent compliance requirements in the market.

And this security goes beyond borders, Abburi says. “Cloud solutions offer tried, tested, and proven security solutions that can be leveraged, ensuring the highest level of security and compliance. Compliance across geographies and jurisdictions is available.”

While a services provider can build applications per stringent compliance needs, firms will still have to adopt best practices and install sound data governance checks. For instance, having a robust disaster recovery plan is critical.

The recent CrowdStrike outage highlighted the folly of not installing a tested disaster management plan, inadvertently offering FX firms an example of what not to do. Carmody says one way of mitigating risk is to avoid overly relying on the public cloud.

“Putting everything in a public cloud, although it offers cost and scale benefits, isn’t necessarily a great business plan for mission-critical trade execution,” he says. “The hybrid model is really what unlocks a lot of advantages for FX trading—reaching the point where one can hyperscale cloud computing in the right way without jeopardising business activity, ensuring that data archived in the cloud is safe and accessible.”

Service providers typically build redundancy into their architecture to ensure local outages do not spread across the network. For instance, data centres connect via networks offering real time data replication. Add the presence of timely backups and firms can expect high availability times.

“Managed services provided by the cloud providers have some of the highest availability ratings, and services are self-healing and redundant,” Abburi says. “Cloud providers cater to holistic observability and monitoring of all aspects of the solution—across infrastructure, platform, and software.”

AI, ML, and the future of the cloud

As with every technological application out there, AI and ML are changing the way firms can interact with their data in the cloud. AI is already a significant part of FX trading strategies, with the proportion of algo-driven trades only increasing.

Integral’s Srivastava says that LLMs have changed the way firms approach large datasets. He cautions that despite their advantages, LLMs and Generative AI are still in their early stages of development.

“Cloud technology—paired with new AI tools—will allow firms to tailor customer experiences even further, and make more advanced data-driven decisions using predictive analytics,” he says. “This will allow businesses of all sizes to innovate and streamline operations in a competitive global market. While foundation models have laid the groundwork, the future of AI lies in the innovative applications built upon them—and there is certainly much more to be seen when pairing these with the cloud.”

While firms can develop in-house LLMs to power trading strategies, deploying these in the cloud comes with a few risks. Carmody identifies two in particular.

“There is risk here in terms of potentially unintended outcomes like a flash crash situation and risk in exposing your trade secret trading strategy,” he says. “There are of course ways to use AI to produce reliable results, but essentially you should still be careful about keeping the underlying data private.”

He gives an example of how IPC mitigates this risk. “IPC’s Unigy platform enables real-time trader voice transcription capture and processing through an AI/NLP engine to provide trading insights to market participants.” He says that inter-dealer brokers are incredibly sensitive about proprietary trading conversations and hosting data in a private cloud works well for them.

“It’s their IP and special sauce, so they might not want to risk running that data through a public AWS Type II or LLM,” Carmody says. “It would be giving away their treasure.”

When asked how he sees the cloud developing for FX use cases, Abburi is positive about the future. “AI and ML will ensure the availability and distribution of applications, in line with increased customer expectations and demands,” he says. “It will also enable alternative channels such as APIs and mobile applications, leveraging cloud platform services to reduce time to market.”

The cloud industry is generally trending towards increased agility and greater innovation. As development accelerates, FX market stakeholders will undoubtedly find plenty of value in deploying to the cloud.