Financial markets have never moved faster.
For over a decade, the industry has optimised relentlessly - compressing latency from milliseconds to microseconds, and now into nanoseconds. Speed has been engineered, scaled, and commoditised.
Yet beneath this progress lies a more important reality: Most firms still do not fully understand - nor control - what is happening inside their trading infrastructure.
They can measure outcomes and observe performance at the surface. Often, however, they cannot explain the behaviour of the systems upon which they depend.
In today's financial markets, this is both a technical limitation and potentially a strategic and regulatory risk.
The illusion of control
Modern trading infrastructure is not just complex, it is also opaque, spanning fragmented markets, distributed systems and layered architectures. Data flows across venues, through networks, into applications, and back out again, with some degree of transformation at every stage. Crucially, much of what matters happens between systems rather than within them.
Latency is introduced across multiple hops. Data is buffered, reshaped, and sometimes lost. Time is synchronised - but not always accurately.
From the outside, things appear to be functioning normally. Systems are up. Trades are executed. Latency is within desired and expected thresholds. All of these good things create the illusion of control.
But underneath the surface lie blind spots, where behaviour is not fully visible, not fully measurable and as a result, not fully understood.
This is the domain of invisible infrastructure - that hidden layer of connectivity, timing and data flow that underpins modern trading performance.
The cost of what you cannot see
The most significant risks in modern markets are not the loud system failures. They are small, transient, incremental and largely unobservable issues:
- A fraction of packet loss during a volatile market event that goes unnoticed, but can result in missing ticks or distorted order books.
- Those few microseconds of unexplained latency that seem negligible, but may be quietly eroding competitive edge in latency-sensitive strategies.
- That almost undetectable microburst of traffic that introduces queueing delays and execution slippage.
Even time itself becomes a source of risk.
In distributed environments, systems rarely align perfectly. Clock drift at a nano or micro-second level makes it difficult to reconstruct a precise sequence of events. In isolation, this may seem inconsequential. Under regulatory scrutiny, it becomes critical.
Increasingly, regulator expectations - and indeed compliance obligations - require firms not just to show what happened, but to be able to demonstrate exactly why and how it happened.
This is where these gaps become material - identifying the elements of a system operating silently and out of sight that may be failing quietly - the "unknown unknowns", to paraphrase an old maxim.
Regulation raising the bar
The shift toward greater observability may be technology-led but it is fuelled by regulation. Frameworks such as DORA, the Digital Operational Resilience Act, are redefining expectations around operational resilience - requiring firms to demonstrate not just the robustness of operational infrastructure, but also measurability, traceability and accountability for what happens within and across it.
At the same time, best execution obligations continue to evolve. What was once just about price efficiency is now a much more nuanced and complex total cost of operation compliance obligation under which regulated firms must be able to show exactly when decisions were made, what data was available at that precise moment in time and how execution paths were determined.
This demands more than logs and summary metrics. It requires granular, time-aligned visibility across the full trade lifecycle.
T+1 settlement adds further pressure to an already bubbling mix, compressing operational timelines and further reducing tolerance for ambiguity. At a time when more and more information is being demanded from them, firms have less time to identify, investigate and resolve discrepancies.
In this environment, infrastructure that cannot be fully observed and explained is not just inefficient. It exposes potential risk - performance and compliance.
Speed is solved. Certainty is not.
Of course, the financial industry continues to talk about speed. Alongside speed, however, is the growing challenge of certainty. Can you explain where any latency comes from? Can you detect transient events occurring in microsecond time frames? Can you reconstruct trading activity with sufficient precision to satisfy regulators and counterparties?
For many firms, the answer to these questions will be a firm "No". Measuring latency is one thing; understanding it is something else entirely, precisely because of the significant portions of infrastructure behaviour that are hidden within layers of abstraction.
But the conversation is shifting:
- From latency to granularity
- From performance to provability
- From speed to certainty
Two firms may operate at similar speeds. But only one can explain, optimise and defend its infrastructure. That is where competitive advantage now lies.
Granularity: the new competitive edge
Latency is a summary metric. Granularity is insight.
Importantly, not all participants require the same level of precision. Some operate in picoseconds, others in seconds. The challenge is that modern trading infrastructure must support both extremes simultaneously - without compromising performance or control.
To understand infrastructure at a meaningful level, firms must move beyond aggregated measurements to packet-level visibility, where individual events can be observed, analysed and correlated across systems. Without this level of granularity, firms cannot isolate cause and effect within complex systems.
In short, it is about making the invisible visible: Microbursts can be detected. Latency can be traced to specific components. Packet loss can be identified in context. Most importantly, behaviour becomes explainable.
This creates both a technical advantage - and a strategic one. In a market where margins are tight and scrutiny is high, the ability to understand systems in detail becomes a prerequisite for performance and control.
Time: the foundation of market trust
If granularity provides visibility, time provides truth. Every aspect of trading is time-dependent. Time is no longer a synchronisation issue. It is a verification issue.
- Sequencing of events
- Validation of execution
- Regulatory reporting
- Dispute resolution
Yet in distributed systems, time is inherently fragile. Clocks drift. Networks introduce delays. Synchronisation is imperfect. Historically, this was tolerated. "Good enough" time was sufficient. That is no longer the case.
As regulatory expectations increase and markets become more complex, time must be precise, constant, traceable and defensible. This is driving the adoption of technologies developed in scientific and metrology environments - where precise time measurement is simply non-negotiable.
For example, the Hadron Collider White Rabbit technology developed at CERN provides sub-nanosecond accuracy and picosecond-level precision in time synchronisation for large-scale distributed systems. Designed originally to synchronise particle accelerator components over long distances, this open-source technology has become a global standard for timing, and is used widely in research, finance and quantum communication.
Specifically, the White Rabbit project ensures that thousands of devices across a network have perfectly synchronised time, with picosecond accuracy, over physical distances. Used in neutrino telescopes, radar applications, and quantum networking experiments, White Rabbit has evident value in financial time synchronisation.
Beyond clock synchronicity, White Rabbit measures and corrects for delay, enabling deterministic time distribution across financial networks with sub-nanosecond precision.
The same level of precision once reserved for quantum physics and scientific research is now becoming a requirement for financial markets. The relevance and value to financial markets is clear. If you cannot prove when something happened, you cannot prove anything about a trade. With precision measurement, time is no longer a background utility - it is an essential layer of control for market integrity.
Measurement: the missing discipline
You cannot optimise what you cannot measure. And most firms cannot measure enough. Alongside timing and control, measurement is another core discipline that remains underdeveloped in many trading environments.
While trading infrastructure is optimised continuously for performance, it is not the case with respect to observability. Firms have built very fast systems, often by layering different technologies and platforms, with incremental enhancements to specific aspects of the infrastructure. Consequently, and collectively, they may not have full knowledge of what all of the moving parts do - alone and in conjunction with other applications.
Without precise, end-to-end measurement:
- Latency cannot be explained fully
- Issues cannot be diagnosed quickly
- Performance cannot be optimised reliably
- Compliance cannot be demonstrated with confidence
Measurement is not just a technical function. It is an operational and strategic capability and differentiator. It shifts firms' operational remit from reactive troubleshooting to proactive control and transforms infrastructure from a black box into a wholly transparent system.
What has long been standard in telecommunications, aerospace and scientific research is fast becoming an imperative in financial markets.
Efficiency, cost, control
While infrastructure complexity increases, cost is a constant constraint. Firms must continuously strike the right balance between investment in performance and operational efficiency and resilience. The prevailing model is shifting:
- Buy commoditised infrastructure where possible
- Build differentiation where it matters
But this introduces its own challenges. Hybrid environments are harder to manage. Visibility is fragmented. Performance becomes harder to optimise. And as stated previously, without measurement, control is largely assumed rather than proven. Firms cannot identify inefficiencies accurately, or validate performance assumptions, or confidently balance costs against outcomes.
In this context, visibility and precision measurement become more than performance-measurement tools. They become strong levers for cost control, governance, and strategic decision-making.
Infrastructure that can prove its own performance
As markets move toward continuous trading, AI-driven execution, and new asset classes, infrastructure requirements will continue to evolve. In short, the next generation of systems must be provable. The question is not how fast can it go, but can you prove what happens at that speed?
Increasingly, firms will need to demonstrate that their systems are deterministic under stress, fully observable across layers, capable of replay and reconstruction and built on the firm foundation of precise, traceable time.
"Unknown unknowns" is attributed most recently to Donald Rumsfeld, former US Secretary of Defense, in 2002, but the phrase echoes an older philosophical maxim and is often quoted in financial markets strategy.
