Tencent Cloud’s Intelligent Traffic Scheduling and High‑Redundancy Architecture Mitigate Shanghai Fiber‑Cut Outage

On the afternoon of March 23, a construction mishap in Shanghai’s Nanhui district cut a fiber optic cable, leading to significant internet service interruptions for many users and businesses, with reports of game disconnections, failed logins, and inability to recharge.

While individual users experienced connectivity issues, some enterprise customers, such as a Tencent Cloud user, saw stable services because the cloud provider’s operations team quickly detected the anomaly and began monitoring the situation.

At 4:52 PM, Tencent Cloud announced that it had observed jitter in Shanghai‑to‑Zhejiang traffic and that the issue was gradually being resolved.

Leveraging its network monitoring platform, Tencent Cloud activated an intelligent traffic scheduling system that redirected public‑internet traffic from Shanghai through its internal T‑level backbone to a Guangzhou‑based telecom exit, then onward to Zhejiang, effectively bypassing the damaged fiber.

The system’s automation allowed it to detect the fault, compute an alternate route, and implement BGP‑based traffic steering without manual intervention, restoring full public‑internet coverage in about two minutes (from 14:40:15 to 14:42:45), with the entire recovery process completing in roughly 150 seconds.

This rapid response is supported by Tencent Cloud’s highly redundant “four‑fiber three‑router” architecture: each availability zone is connected by three independent fiber links from different directions and two physically isolated wavelength‑division multiplexing (WDM) systems, enabling sub‑50 ms automatic failover when a fiber is cut.

Additionally, the network incorporates optical time‑domain reflectometers (OTDR) that can pinpoint the exact location of a cable break, providing precise data to carriers for swift repair.

The case demonstrates how advanced cloud networking technologies—automated traffic engineering, high‑availability design, and real‑time fault detection—can maintain service continuity even during severe infrastructure failures.