The 300 kilometers per hour-Testbed

What Racing Teaches Telecom Before Everyone Else Finds Out

In 1951, an American privateer named Briggs Cunningham experimented with a radio link between his pit and his car at the 24 Hours of Le Mans. It did not catch on.

For years afterward, one of the fastest reliable ways to tell a driver anything was still a person holding up a board. At a circuit as long as Le Mans, even that was not enough: teams stationed signallers at Mulsanne, connected to the pits by telephone, so a decision made as the car passed could be relayed, written onto a board and shown to the driver several minutes later.

Two to three minutes of latency, delivered by hand.

We start here because much of the history of mobile connectivity is hiding inside that story. Racing has repeatedly pushed wireless systems toward conditions that ordinary networks encounter only later: mobility, interference, strict power budgets, dense temporary infrastructure and an unforgiving requirement for reliability.

The racetrack has been a communications testbed for decades. What makes it interesting today is not that racing has somehow solved the future of telecom. It is that the constraints appearing there increasingly resemble the ones emerging elsewhere.

A century of catching up

The evolution of racing communication reads like a compressed timeline of the mobile network itself.

The board eventually gave way to CB and FM radio in the 1970s. By 1977, teams including Alpine and Porsche were equipping cars with radio systems. Small aerials appeared on the vehicles; larger antennas appeared around the paddock because the signal still had to reach distant parts of circuits through terrain, structures and vegetation that were not designed with radio propagation in mind.

That last detail matters.

A race car combines several difficult wireless conditions at once. It is moving rapidly. Its radio environment changes continuously. The vehicle imposes tight restrictions on weight, power and packaging. Metal structures, barriers, buildings and other transmitters complicate propagation. And the value of information falls rapidly when it arrives late.

None of these problems belongs uniquely to motorsport.

They reappear in connected mobility, industrial networks, logistics sites, autonomous systems and increasingly in dense private-network deployments. Racing simply concentrates them into one place and forces the infrastructure to cope under unusually visible conditions.

That makes the circuit useful not as a blueprint, but as a stress test.

The scale nobody sees coming

Formula 1 illustrates how far this has progressed.

A current F1 car carries more than 300 sensors capable of generating over a million telemetry data points per second while it is running. Across a race weekend, Formula 1 processes several terabytes of telemetry before adding video, audio, timing data and the rest of the broadcast operation.

At the level of the overall event, the numbers become much larger. Around 650 terabytes of data can move through Formula 1's technical and broadcast infrastructure during a race weekend.

The architecture behind that figure is arguably more interesting than the number itself.

A temporary technical environment has to be deployed at a circuit, connected to remote production and engineering facilities, operated under heavy load and then dismantled and reproduced at another venue. Compute increasingly moves closer to where the data is created so that information does not have to travel unnecessarily through the entire infrastructure before it can be processed.

In other words, the problem is no longer simply connectivity.

It is connectivity plus edge computing, transport, orchestration, resilience and application requirements—all assembled as one system.

That begins to look familiar to anyone working on private 5G, industrial networks or future 6G architectures.

The real problem is getting it off the car

The glamorous version of the problem is “5G at the Grand Prix.”

The harder engineering question is what happens between the sensor and somewhere useful.

Cars generate telemetry continuously. Cameras add bandwidth-hungry video. Trackside systems create their own streams. New production techniques introduce fast-moving cameras and drones. Experiments with millimetre-wave communications have demonstrated that very high-capacity wireless links can also be maintained in demanding motorsport environments.

But high peak throughput is only one part of the problem.

A useful system must manage mobility, handover, coverage, interference, backhaul, compute placement and failure simultaneously. Improvements in one layer can simply expose a bottleneck in another.

That distinction matters outside racing because modern networks are becoming less symmetrical in their assumptions.

For most of the consumer-mobile era, network planning has concentrated heavily on the downlink. People consumed webpages, music and video, so capacity largely flowed toward the device.

The emerging edge looks different.

Cameras upload continuously. Industrial sensors generate streams. Vehicles produce telemetry. AI applications increasingly capture images, video and contextual information locally before sending data toward compute resources elsewhere.

Measurements published by Ericsson illustrate the direction of travel: among 55 service providers studied, 43 experienced faster growth in uplink traffic than in downlink traffic. Downlink still dominates total volume, but the growth pattern is changing.

That distinction is important. We are not suddenly entering an entirely uplink-dominated internet. We are entering a world in which network architectures designed around downlink dominance have to accommodate a much wider variety of traffic patterns.

Racing has lived with that problem for a long time.

Telemetry is predominantly about getting information off the vehicle. Formula 1 also deliberately limits what can happen in the opposite direction: pit-to-car telemetry is prohibited, preventing teams from remotely controlling or reconfiguring the car through a telemetry channel.

The interesting lesson is therefore not simply that racing is “uplink-heavy.”

It is that future networks may increasingly need to distinguish between data collection, control, human communication and machine actuation—and apply different reliability, latency and security rules to each.

The network around the network

Once you look at the circuit this way, the car is only one piece of the architecture.

A temporary event may require high-capacity access networks, front- and backhaul connections, local compute, redundant transport and mechanisms for distributing the same content to very large audiences.

Running fibre everywhere is one answer, but not always the only one. Temporary or difficult-to-reach locations create room for complementary technologies such as optical wireless backhaul. Dense zones can require locally deployed small cells or private mobile infrastructure. One-to-many traffic such as live video raises another architectural question: whether continuously reproducing identical unicast streams is always the most efficient way to move the same information toward many recipients.

These are different layers of the same system.

In a modular architecture, they do not have to come from one vertically integrated supplier.

An optical link can provide transport. A small-cell or private-network implementation can provide radio access. Multicast mechanisms can change how content is distributed farther up the stack. Edge compute can be introduced or moved independently as application requirements change.

Projects such as fiber-like, airpuls and QUICast illustrate these three different positions in the stack: optical wireless transport, configurable private-radio infrastructure and multicast-oriented content delivery respectively.

The interesting point is not the individual companies. Other implementations can occupy the same positions.

It is the architecture that appears when those positions are treated as interchangeable building blocks.

The part racing cannot answer for us

There is a temptation with testbeds to look for a technology that “wins.”

Racing suggests almost the opposite conclusion.

The communications problem keeps changing.

Radio solved the signalling problem but created coverage and reliability questions. More telemetry created transport problems. More cameras created bandwidth problems. Remote production moved processing elsewhere and made resilient connectivity still more important. Edge computing moves some of that processing back toward the circuit.

Every solved bottleneck reveals another one.

That is why openness matters.

Private 5G and future 6G networks will probably have to operate for much longer than any individual hardware generation or software release. Operators of factories, ports, hospitals, research facilities and transport systems therefore have an interest not only in what a network can do on the day it is installed, but in whether parts of it can be replaced, extended and independently tested several years later.

Open interfaces and modular RAN architectures do not automatically guarantee that outcome. They introduce their own integration, testing and operational complexity.

But they make a different model possible: one in which radio, transport, compute and applications can evolve at different speeds rather than requiring an entire system to be replaced whenever one layer changes.

That is especially relevant in Europe, where private networks and research testbeds can serve two purposes simultaneously. They can support immediate industrial applications while also giving researchers, startups, equipment suppliers and operators somewhere to test components against real systems rather than only against simulations or proprietary reference platforms.

The strategic question is therefore larger than whether Europe can produce another complete telecom vendor.

It is whether Europe can sustain enough interoperable components, implementation knowledge, test infrastructure and system-level expertise to keep meaningful choices open as networks evolve toward 6G.

The 300 kilometers per hour testbed

A racetrack is useful because it compresses the problem.

Mobility is faster. Failure is more visible. Infrastructure is temporary. Data volumes are high. Radio conditions are difficult. Decisions have to happen quickly.

What looks extreme there often looks less unusual a few years later.

Connected vehicles, autonomous machines, AI-enabled cameras, industrial sensing and distributed compute are gradually bringing some of the same traffic patterns into ordinary networks.

The lesson from racing is therefore not that telecom should copy Formula 1.

It is that the future network is unlikely to be one network at all.

It will be radio access, transport, compute, control and applications assembled around a particular environment—and then changed when that environment changes.

At speeds of up to 300 km/h that becomes obvious earlier than almost anywhere else.

Weiter
Weiter

From Customer Discovery to Sales: How to Turn Interest into Evidence (Pt 2)