Official account of Tencent Cloud Middleware. Focuses on microservices, messaging middleware and other cloud‑native technology trends, publishing product updates, case studies, and technical insights. Regularly hosts tech salons to share effective solutions.
Tencent Cloud's MQTT service now commercially supports MQTT over QUIC, leveraging QUIC's connection migration, 0-RTT handshakes, and stream-level multiplexing to drastically reduce device disconnection windows and server reconnection storms in mobile IoT scenarios like connected vehicles and drones.
Enterprises deploying large-model APIs often struggle to see token usage, latency, and errors; the AI Gateway embeds metrics, structured logs, and distributed tracing at the gateway layer, providing token-level insights, request-level latency breakdowns, and full-chain auditability without code changes, as demonstrated in a real-world incident.
This article dissects TDMQ RocketMQ's Lite Topic, explaining its two‑level topic model, automatic lifecycle, storage reuse, client‑level subscription management with Ready Event Sets, and the exclusive "latest‑connection‑wins" consumption mode that together enable millions of lightweight channels for AI‑native applications.
A CTO discovers a three‑fold surge in AI model token bills with no clear usage breakdown, prompting a deep dive into Tencent Cloud's AI Gateway, which offers multi‑dimensional quota governance, real‑time cost monitoring, semantic caching, and a unified cost‑management dashboard to make every token expense transparent and controllable.
The article analyzes the communication challenges of AI‑native applications—async coordination, per‑session isolation, and millions of independent channels—and shows how RocketMQ Lite Topic solves them with lightweight, auto‑reclaimed topics that guarantee strict ordering, fault‑tolerant streaming, and seamless break‑resume handling.
The article explains how Tencent Cloud’s AI Gateway uses intelligent routing and a three‑layer high‑availability design to distribute traffic, handle model failures, and support multi‑model, multi‑vendor, and compliance‑driven scenarios, ensuring large‑model services remain stable and continuously available in production.
Enterprises with hundreds of RESTful APIs often cannot expose them to AI agents, not because models are weak but because tool integration is cumbersome; the article shows how the MCP Gateway enables zero‑code conversion of existing interfaces, centralizes authentication, logging, and protocol translation, and supports multi‑backend scenarios for seamless AI ecosystem access.
The article analyzes why large‑scale delayed messaging is needed, identifies the bottlenecks of the Apache Pulsar community solution, and explains TDMQ Pulsar's three‑step redesign—hierarchical time wheels, expiration re‑push, and immutable message IDs—that together enable stable million‑message delayed queues with controlled memory and minute‑level hole impact.
This article explains the challenges of traditional point‑to‑point AI agent communication, introduces the A2A protocol and its limitations, and details how combining A2A with MQTT via Tencent Cloud TDMQ creates a dynamic, loosely‑coupled, and scalable solution with practical SDK examples and real‑world case studies.
The article explains how RocketMQ 5.x’s storage‑compute decoupling and POP consumption model require a new gRPC client, and how Tencent Cloud’s virtual‑queue solution provides full compatibility for legacy Remoting SDKs by abstracting queues and transparently converting consumption modes, eliminating client‑side rebalance and preserving order guarantees.
