The following section covers the concepts and data flow from data source to the Vessel Insight cloud storage.
End-to-end flow
Vessel Insight is a data infrastructure for:
- Collecting raw sensor data from onboard sources on vessels,
- Sending data to cloud where it is contextualized,
- Share data with specialized applications or other data consumers.
At a high-level, the flow is as shown in the following diagram:
The Data Sources are systems that generate Data Points. The data points are produced by what we refer to as Sensors in the data source system, often called tag or a signal in a source system setting. The data sources are interfaced to an IoT Gateway, the Vessel Insight Edge (edge computer) via various protocols and architectures. Commonly used are MQTT interface connectors on history stations, NMEA or MODBUS serial interfaces. The edge computer runs IoT software components that is managed from the Kongsberg Digital cloud using our Secure Developer Lifecycle process. The edge can forward data to containerized apps, other onboard consumers and to cloud. The edge also fetches the tag metadata, such as unit. Data is compressed in packages and sent to cloud via the secured vessel communication carriers and stored in a cloud database system. The sensors that are selected to be sent to cloud is referred to as Replicated.
To reduce the complexity of integrating with the data, a context layer is added to the raw sensor names. This is done by mapping the sensors to a structure called the Asset Model, which is a vessel data model where standard Paths represents typical system properties. Each sensor is Mapped to the relevant path, and the mapping is kept consistent across the fleet.
Cloud-side applications can pull data from Vessel Insight cloud storage via a REST API or receive the data stream directly via IoT streaming interfaces.
Key takeaways:
- Raw data is collected as is from the source systems.
- If the sources do not transmit data, there will be no data in the cloud for the applications.
Connectivity loss situations
When internet connectivity is lost, sensor data is temporarily stored in a buffer on the edge computer's disk. This buffer is managed through various configurable options, including size, number of messages, and time. It is preset to a default configuration suitable for all vessels.
Data Transfer delay
The transfer delay of data from sources to cloud storage can vary based on the data source protocol type, software configurations, and communication latency and noise. However, consumers typically see data appearing in the REST APIs within minutes after it is captured on vessels.
