Introduction to TencentOS tiny: Open‑Source IoT RTOS, Solutions and Development Practices

TencentOS tiny is an open‑source, lightweight real‑time operating system (RTOS) designed for the Internet of Things (IoT). It was officially open‑sourced on September 18 and quickly rose to the second place on GitHub’s hot list, gathering more than 3,500 stars and 800 forks. The project supports over 50 hardware platforms and integrates common IoT protocols such as MQTT, CoAP and LoRaWAN.

The OS provides a minimal kernel with task management, memory management, IPC, and exception handling, as well as a rich set of IoT components including low‑power frameworks, OTA upgrades, key‑value storage, and debugging toolchains. Its advantages are small footprint, low power consumption, extensive IoT components, a reliable security framework, good portability, and convenient debugging.

Two typical solution cases are presented:

Smart cabinet: TencentOS tiny controls the cabinet, integrates with a central system and cloud AI to achieve scan‑to‑open, item retrieval and automatic settlement. Gravity sensing is used to improve AI recognition when items are occluded.

Smart agriculture: Sensors collect temperature, humidity, pH, oxygen, and light intensity, upload the data to an IoT cloud platform, and the cloud AI issues greenhouse adjustment commands (e.g., turning on supplemental lighting).

A hands‑on development practice is described. A custom development board based on STM32L4 is used, with interchangeable sensor modules (temperature/humidity, light) and communication modules (NB‑IoT, 2G, Wi‑Fi). The startup flow includes board‑level initialization, kernel initialization, task creation, and kernel start‑up. Task creation is simple: configure priority, stack size, and optionally use the CMSIS API.

The importance of an RTOS is explained: in time‑critical scenarios (e.g., low‑light lighting control or smoke‑sensor fire alarm) high‑priority tasks pre‑empt lower‑priority ones, guaranteeing real‑time response. TencentOS tiny uses a priority‑preemptive scheduling strategy, with optional time‑slice scheduling, and provides mutexes, semaphores, events, and queues to simplify concurrent programming.

For network connectivity, the board uses a communication module controlled via AT commands. TencentOS tiny offers an AT‑adaptation framework that abstracts module differences and presents a socket‑like API (connect, send, recv) to the application layer, making it easy to replace modules without changing business logic.

Data is uploaded to the cloud through integrated IoT protocols and the Tencent Cloud SDK. The platform supports device management, device shadows, and rule‑engine forwarding to databases, message queues, or big‑data services. A mini‑program (WeChat) is used as the end‑user interface, leveraging built‑in Bluetooth, Wi‑Fi, and NFC capabilities to interact with devices. Cloud development features (cloud database, storage, functions) reduce development cost and eliminate domain registration and备案.

The final practice combines sensor data acquisition, real‑time control, MQTT communication with Tencent IoT Hub, and mini‑program cloud functions for querying and controlling devices. The whole solution demonstrates a low‑cost, fast‑to‑market development model for startups.

Q&A: The speaker confirms that developers can use the provided templates to build other solutions, highlights the easy portability of TencentOS tiny, and notes that with basic embedded experience the kernel can be ported within half an hour using the detailed documentation on GitHub.