Core principles
The nRF Asset Tracker is built on the following principles:
- Teach by showing
All the examples are designed to solve a concrete use case (an asset tracker) instead of providing generic or abstract solutions. It is not a framework, but a real application.
- Err on the side of security
It follows the most robust security recommendations of the respective cloud provider.
- Single tenancy
It implements a scenario in which all the authenticated users can be trusted to access all the devices. This is a typical scenario for cellular IoT products, and it simplifies the onboarding of new devices.
- Serverless
It uses a serverless architecture as much as possible to have near-zero costs for the operation during the development, and to provide horizontal scaling of resources to be used in a production system if needed.
- Cloud native
It follows the best practices of the respective cloud provider to reduce development efforts due to abstraction and offer the best integration into the cloud provider’s ecosystem.
- Being offline is not an exception
Highly-mobile cellular IoT products need to handle unreliable connections gracefully, by implementing the mechanisms to retry the failed sending of data. This also means that the sensor measurements need to be timestamped when they are created, and not when they arrive at the cloud.
- Maximize power saving
The firmware examples highlight the power saving features of the nRF9160 since this is a critical aspect for developing small form-factor devices that need to run on the battery charge of a single battery cell for years.
Power saving is used as much as possible in all the operating states of the device.
All the necessary computation that can be done on the device must be done on the device itself.
The least amount of data is sent to the cloud (optimize for message size).
Data is encoded in the most efficient way for the device (optimize for computation cycles).