Customer Story – Marc Willems, Founder, Bitchange
Marc Willems’s goal wasn’t to build a trading bot. He set out to see the market as clearly as possible.
That distinction matters more than it sounds. Most crypto tools show you what already happened. Marc wanted to watch it happen – price changes, trades, liquidity shifts, order books moving in real time across dozens of exchanges simultaneously. Before any algorithm, before any signal, before any automation: just a clean, honest picture of what the market is actually doing.
Building that picture turned out to be a project in itself.
The problem with crypto data
Crypto markets are fragmented in a way that’s easy to underestimate. It’s not one market: it’s hundreds of exchanges, spread across different regions, running different APIs, quoting in different currencies, behaving differently depending on where in the world you’re connecting from.
If you want a serious view of what’s happening, you cannot rely on one exchange or one data source. The moment you centralise, you start seeing a partial picture – filtered through one location’s latency, one uplink’s capacity, one region’s network quirks.
Marc’s answer to this was TRON: a distributed market intelligence engine that collects and normalises live data from roughly 30 exchanges, tracking more than 18,000 instruments. During active periods it processes tens of thousands of market events per second – ticks, trades, order book changes – and turns them into one coherent view of the market.
On top of that foundation, Marc builds archives, visualisations and research tools. Automated trading comes later, but the data layer has to come first.
“I do not believe in starting with the bot. The market data foundation has to come first. TRON is that foundation.”
The plan that didn’t survive the reality check
The original infrastructure idea was straightforward: a few large, powerful dedicated machines handling most of the compute centrally. Big CPU, lots of memory, everything in one place.
Then prices moved. Hardware and hosting costs for that kind of setup went up significantly – in some cases roughly 3x compared to what Marc had budgeted. For an independent developer without a company behind him, that changes the equation completely.
So the plan changed. Instead of scaling vertically with a handful of behemoth machines, Marc started thinking horizontally: multiple regional VPS nodes, each doing part of the work, placed closer to the exchanges they’re watching.
It turned out to be a better fit for the project anyway.
Why regional infrastructure matters (and what happens without it)
For crypto market data, milliseconds are not vanity numbers. They affect how fresh your picture of the market is.
When everything runs from one location, several things happen at once. Latency to distant exchanges gets worse. API response times become uneven. One network issue affects the whole system. One uplink becomes a bottleneck. And when you need to scale, you can only go up – more hardware in the same place.
With regional nodes, you can put collectors close to the exchanges they’re watching. You can split workloads. You can compare which regions perform best for which exchanges and route accordingly. And when something needs to grow, you add a node where it makes sense rather than upgrading everything at once.
Marc now runs multiple Contabo VPS nodes across Europe, the UK, Japan, Singapore and the United States. Some handle exchange data collection. Some manage monitoring and routing. Some are used for testing new parts of the system. The logic is simple: let each node do the work it’s best positioned to do, then feed the results back into TRON.
Why Contabo
Two things according to Marc: price-performance and regional availability.
After the cost of large dedicated systems stopped making sense, Marc needed infrastructure that would let him scale horizontally without the bill becoming the main problem. Contabo made that possible.
“I do not necessarily need one giant machine. I need several reliable machines in different regions, with enough CPU and memory to run collectors and processing tasks continuously. Contabo makes that affordable. That matters a lot when you are an independent developer building a serious infrastructure project without a large company budget behind it.”
There’s also a practical dimension that’s easy to overlook. With distributed crypto data collection, you can’t judge everything from a spec sheet. You have to run real workloads, measure actual latency, observe stability over time, and see how each node behaves under live market conditions. That process takes time, and it requires infrastructure you can actually afford to experiment with.
What changed
The biggest shift wasn’t a number – it was architectural.
Instead of pushing all data collection through one fixed location, Marc can now run regional workers and compare performance per region. Response times improved for several exchanges. Pressure on the central infrastructure dropped. And the way he thinks about scaling changed: not bigger, but more distributed.
The system currently tracks around 30 exchanges and more than 18,000 instruments. It processes tens of thousands of market events per second during active periods. That’s running across a set of regional VPS nodes that, individually, are modest – but together, positioned well, do serious work.
Marc has also made a public demo available at https://x64now.com – a simplified live view of part of the market data layer, running on one of the Contabo nodes. It does not show the full TRON system, but it gives a simplified public view of part of the live market data layer TRON produces.
What’s next
The live data layer is running. The next priority is the historical and analytical layer: solid archives, timeframes, transaction data, order book history, replay tools.
After that, signal research. Then, eventually, automated execution.
Marc describes it as a layered process: observe the market correctly, store and replay history reliably, identify patterns, build signals, and only then connect to automated trading. Each step depends on the one before it.
TRON is the foundation. The rest gets built on top.
If you’re an independent developer building serious infrastructure and want compute that doesn’t make cost the main constraint, take a look at our VPS plans – and explore the public demo at https://x64now.com.
Want to be featured? We’re always looking for customers doing interesting things on Contabo infrastructure. Get in touch.
