Tourisme Participations runs 15 electric cruise boats on the lakes of the French-Swiss Alps. AnotherTrail rebuilt fleet operations on EdgeLake — onboard AI, decentralized SQL, and zero cloud dependency.
Tourisme Participations runs 15 boats across the three main lakes of the French-Swiss Alps, cycling between commuter routes and tourist cruises since 2010. The CEO's transition from diesel to electric propulsion needed a data layer that could match the long lifecycle of a ship (major overhauls every 25 years) without locking the fleet into a cloud vendor. Their partner, AnotherTrail, picked EdgeLake.
Each boat carries an ARM-based industrial PC running Linux and a containerized EdgeLake-AnyLog operator node. The operator aggregates data from the Torqeedo propulsion system, the energy management system, the chargers, the inverters, the generator, and the engines, all over Modbus, with Python scripts handling the few non-standard interfaces. A shore-based master node coordinates metadata, and queries route directly to the relevant boat.
The structural argument for staying on edge is bandwidth and cost. Streaming raw telemetry from boats in the middle of an Alpine lake over 4G is both expensive and unreliable. EdgeLake's pattern of letting queries traverse the network instead of raw data means deployments measured in days rather than months, no per-vessel cloud setup, and the cloud is used only for archival and redundancy at storage prices instead of data-service prices.
The planned next phase mirrors the WINNIIO trajectory. AnotherTrail is layering on onboard AI inference for battery-life prediction and route optimization, plus federated learning so each ship contributes to a shared model without exporting its raw operational data. The fleet learns as one without losing data sovereignty per vessel.