Challenges and Exploration of IoT Technology in Industry Applications

This article is a transcript of Chen Jie’s presentation at the Techo TVP Developer Summit, titled “From Connectivity to Intelligence – Challenges and Exploration of IoT Technology in Industry Applications.” It provides a comprehensive overview of the development history, system architecture, product capabilities, and practical case studies of Tencent Cloud IoT.

1. IoT History and System Layers – The speaker divides IoT evolution into three stages: the early “萌芽期” (late 1990s‑2005), the initial growth period (2005‑2014) marked by the emergence of consumer devices and the standardization of MQTT, and the rapid expansion phase (2015‑present) driven by cloud‑based IoT platforms. The IoT stack is described in four layers: Device Perception, Transmission, Platform, and Application.

2. Tencent IoT Product Capabilities and Service Architecture – Tencent’s IoT journey started with the “QQ物联” product (2013‑2014) and matured into a unified communication platform that supports billions of devices and millions of QPS. The platform consists of three core communication services (narrow‑band, broadband, and edge‑node management), a developer platform offering SDKs, data‑visualization, and SaaS hosting, and a suite of security services (high‑defense, one‑device‑one‑secret, TLS/DTLS, TID). The architecture separates connection handling (Conn‑proxy) from business logic (Conn‑logic) using in‑memory message queues for upstream and downstream traffic, enabling high reliability and overload protection.

3. Industry Scenario – Consumer Speaker – The speaker examines a large‑scale consumer speaker deployment, highlighting challenges such as massive data volume, heterogeneous hardware resources, and weak‑network environments. Solutions include multi‑region deployment, Http‑DNS with fallback mechanisms, near‑edge node selection, and robust message retry/queueing strategies to achieve sub‑second latency even on 2G/4G networks.

4. Reliability and Security Measures – Reliability is ensured through multi‑region redundancy, 1‑master‑2‑standby database (TDSQL) with cross‑region sync, and KV cache synchronization via global MQ. Security is addressed with DDoS protection, device‑level encryption (one‑device‑one‑secret, soft‑hardening), TLS/DTLS/TID protocols, and fine‑grained monitoring and alerting.

5. General Access Capabilities – To lower the entry barrier for traditional hardware manufacturers, Tencent provides a four‑layer SDK architecture (Hardware, Network, IoT Business, User Business), unified testing standards, protocol adapters (e.g., Bluetooth, Wi‑Fi), and zero‑code Wi‑Fi provisioning methods (SoftAP, Bluetooth‑assisted, one‑click). SaaS capabilities allow developers to build applications via drag‑and‑drop APIs.

6. Audio‑Video Scenarios – The presentation describes three audio‑video solutions: TRTC (cloud‑mediated conferencing), XP2P (P2P‑based low‑cost streaming), and CSS (traditional live‑streaming). It explains how IoT signaling (MQTT) is used for device discovery and connection establishment, achieving high success rates and low latency for real‑time audio‑video communication.

7. Speaker Introduction – Chen Jie is a Tencent Cloud IoT expert engineer with extensive experience in designing and implementing large‑scale IoT solutions.

8. Recommended Reading – Links to related articles on edge computing, 5G IoT evolution, TencentOS for AIoT, and tracing technologies are provided.