For the complete documentation index, see llms.txt. This page is also available as Markdown.

Overview

Cloud Networking is a private, on-demand, full-mesh global dedicated network. It bypasses the congested public internet and connects your data centers, offices, and public clouds around the world into one unified "intranet", making cross-region and cross-cloud data transmission low-latency, low-loss, and predictable.

Cloud Networking Overview

What Problems Does It Solve

Cloud Networking is built for you if you need any of the following:

  • Cross-region interconnection — multiple data centers / branch offices located in different countries that need stable, low-latency connectivity;

  • Hybrid / multi-cloud interconnection — on-premises IDCs or offices that need private direct connections to public clouds such as AWS, Google Cloud, Alibaba Cloud, and Tencent Cloud, without going over the public internet;

  • Performance-sensitive workloads — latency- and jitter-sensitive scenarios such as gaming, financial trading, real-time audio/video, and core enterprise systems;

  • Fast provisioning and elastic adjustment — provision on demand and adjust bandwidth as needed, avoiding the weeks-long delivery cycles of traditional leased lines.

Technical Architecture

Cloud Networking uses SR (Segment Routing) as its underlying protocol to build an all-new global backbone. Designed for the SDN architecture, SR runs on a fully in-house controller with complete separation of the control plane and the forwarding plane — this prevents service impact caused by controller failures, and when a backbone link fails, the controller can deliver service paths according to specific rules, providing superior resilience, flexibility, path selection, and network telemetry for advanced traffic engineering.

Last updated