Background
SailADV is revolutionizing how luxury yachts are monitored and maintained. As a leader in maritime engineering and testing, SailADV works closely with shipyards and OEMs to deliver high-performance telemetry and predictive diagnostics platforms that track over 1,000 individual metrics from every yacht, from propulsion systems to HVAC units and energy performance.
With a growing customer base and plans to scale from dozens to hundreds of yacht deployments, SailADV needed a new data backbone that was fast, lightweight, scalable, and easy to maintain across both onboard and cloud environments.
The Challenge
Before adopting TDengine, SailADV was using PostgreSQL with the TimescaleDB extension. While the relational model offered flexibility, it quickly became a bottleneck:
- Performance: Query latency for cloud analytics grew to 50 seconds on a 90GB dataset.
- Storage: The relational schema required storing complex JSON blobs or maintaining separate rows per metric.
- Dual-stack complexity: Maintaining separate edge and cloud stacks slowed development.
- Scaling: Needed a solution for scalable, efficient data ingestion and long-term storage.
“We were looking for a database that we could use both onboard and in the cloud, without incurring huge maintenance costs.”
— J.G., CTO, SailADV
Why TDengine
| Feature | TDengine | Other TSDBs (e.g. TimescaleDB) |
|---|---|---|
| Edge-cloud unification | Seamless deployment on both | Different stacks required |
| Data model | One table per metric for fast, flat schema | Complex joins and nested JSON |
| Query performance | Under 1 second for 3+ months of data | Up to 50 seconds |
| Storage efficiency | Automatic compression | High storage cost with JSON |
| Lightweight footprint | Low memory and CPU usage | Heavy compute required |
TDengine’s “one table per metric” model aligned naturally with SailADV’s architecture, enabling fast access to specific signals while storing timestamps, values, and metadata as flat records.
“The ingestion and query performance is extremely fast; querying three months of data for a single variable now takes less than one second.”
— Sergey Soltan, Software Engineer, SailADV
Architecture and Implementation
SailADV’s telemetry system collects over 1,000 variables from each yacht, including temperature readings, fan settings, and energy metrics. These are aggregated using industrial protocols and integrated into TDengine at the edge and in the cloud.
- Edge setup: Local TDengine ensures offline collection and Kubernetes ensures high availability.
- Cloud sync: Data is synced to AWS for analytics and customer access.
- Data modeling: Each variable stored in its own subtable using a shared supertable.
“We designed everything to scale, from unified schema to deployment scripts.”
— J.G., CTO, SailADV
Results
- Performance: Query time dropped from 50 seconds to under 1 second.
- Storage: Compression cut telemetry export sizes significantly.
- Simplicity: Unified architecture reduced development and maintenance time.
- Future-proofing: Built-in retention and flat schema simplify scale-up.
Looking Ahead
SailADV plans to roll out predictive diagnostics and condition-based maintenance features by late 2025.
Because of the way SailADV organizes and labels its data — each metric enriched with metadata and system relationships — they are well-positioned to apply anomaly detection, fault correlation, and root cause analysis. They’ve also built an adjacent metadata database that links each variable to a component, subsystem, and vessel, enabling rich contextual analysis.
“TDengine is not just a database — it’s the foundation for the next generation of intelligent marine systems that we’re building.”
— Giovanni Palamà, CEO, SailADV
Summary
SailADV’s transition to TDengine enabled them to unify their edge–cloud stack, simplify their data model, and dramatically boost query performance. With ambitions to scale their intelligent telemetry platform to hundreds of vessels, SailADV sees TDengine as a long-term partner for building cutting-edge marine analytics and diagnostics.
