Cloud Computing 11 min read

Optimizing Hybrid Cloud Bandwidth Utilization with Tencent Cloud Connect Network: A Case Study of Enterprise A

Enterprise A consolidated its numerous VPC peering links and dedicated‑line gateways into a single Tencent Cloud Connect Network instance, enabling automatic route learning, ECMP load‑balancing across four 10 Gbps lines, boosting bandwidth utilization from about 50 % to 75 % while cutting operational complexity and enhancing cross‑region high availability.

Tencent Cloud Developer
Tencent Cloud Developer
Tencent Cloud Developer
Optimizing Hybrid Cloud Bandwidth Utilization with Tencent Cloud Connect Network: A Case Study of Enterprise A

Enterprise A, a large internet company, operates a hybrid‑cloud environment with nearly 100,000 CPU cores in the cloud and more than 24 Gbps of dedicated line traffic in the East China region. Its resources are primarily deployed in East China, with additional zones in North China, South China, and Southwest China. The company seeks to improve bandwidth utilization by 50% using a cloud networking architecture.

1. Peering Connection Overview – Peering connections provide high‑bandwidth, high‑quality inter‑VPC communication across multiple regions, accounts, and heterogeneous networks. They support VPC‑to‑VPC and VPC‑to‑Blackstone private network connectivity.

2. Traditional Dedicated Line Overview – In the initial hybrid‑cloud setup, dedicated lines connect the on‑premise IDC to cloud VPCs. Multiple VLAN IDs are required when a single physical line must serve several VPCs, leading to complex management.

3. Cloud Connect Network (CCN) Overview – CCN enables inter‑VPC and VPC‑to‑IDC connectivity with automatic route learning, multi‑point interconnection, link selection, and rapid fault convergence. It supports over 100 Gbps bandwidth and 99.99% availability across more than 20 regions.

Enterprise A Architecture

• The company has two VPCs in Shanghai (business and test) and additional VPCs in Beijing, Chengdu, Guangzhou, and Hong Kong. All VPCs are linked via peering connections and traditional dedicated lines to the on‑premise IDC.

• Peering connections in the same region theoretically have unlimited bandwidth, limited only by the underlying data‑center DCI capacity.

• Four dedicated 10 Gbps lines connect the IDC to the cloud, each carrying about 6 Gbps (25% of total traffic). Current utilization exceeds 60%, causing congestion when a line fails.

Challenges of the Existing Interconnect

Maintaining a large number of peering connections and routes as VPC count grows increases operational cost.

Cross‑region peering gateway failures do not automatically switch routes, leading to prolonged outage.

Ensuring high availability when any one of the four IDC‑to‑cloud lines fails requires complex BGP and gateway configurations.

Cloud Connect Network Refactoring Solution

1. VPC‑to‑VPC Interconnection via CCN – All VPCs join a single CCN instance, which automatically learns routes, eliminating the need for numerous peering connections and manual route entries.

2. VPC‑to‑IDC Interconnection via CCN – The four dedicated‑line gateways also join the same CCN instance. Traffic from any VPC is routed to the CCN, which performs ECMP load‑balancing across the four gateways. If a physical line fails, the remaining gateways share the load, maintaining service continuity.

3. Benefits include reduced operational overhead, higher cross‑region availability through multi‑AZ gateway clusters, and increased bandwidth utilization from ~50% to ~75% without additional cost.

Key Takeaways

Joining all VPCs to a CCN eliminates the need for pairwise peering, lowering maintenance effort.

Multi‑AZ gateway clusters improve resilience and availability.

Automatic traffic redistribution upon line failure enhances high‑availability of IDC‑to‑cloud links.

Only four BGP peers are required for all VPC‑to‑IDC connections, dramatically reducing dedicated‑line operational costs.

Implementation Considerations

• Conduct thorough pre‑migration analysis (business characteristics, bandwidth requirements, current architecture shortcomings, migration difficulty per region).

• Follow a phased gray‑deployment: test with a single VPC, then gradually scale based on traffic volume.

• Verify gateway capacity, MTU, and CCN throttling algorithms (distributed algorithm recommended for optimal bandwidth usage).

• CCN limits: up to 5 instances per account, each instance can bind up to 25 network instances (e.g., VPCs). Bandwidth throttling can be configured per region or per inter‑region link.

Reference documents: VPC test configuration, peering‑to‑CCN migration guide, traditional dedicated‑line‑to‑CCN migration guide.

network architecturecloud computinghybrid-cloudbandwidth optimizationTencent CloudCloud Connect Network
Tencent Cloud Developer
Written by

Tencent Cloud Developer

Official Tencent Cloud community account that brings together developers, shares practical tech insights, and fosters an influential tech exchange community.

0 followers
Reader feedback

How this landed with the community

login Sign in to like

Rate this article

Was this worth your time?

Sign in to rate
Discussion

0 Comments

Thoughtful readers leave field notes, pushback, and hard-won operational detail here.