Business Continuity Solutions on Tencent Cloud: High Availability and Disaster Recovery

Business Continuity (BC) refers to an organization’s ability to continue delivering products or services at an acceptable level after an interruption. It identifies critical functions, potential threats, and adopts technical measures, plans, and processes to ensure those functions remain operational under any circumstances.

BC scenarios include general outages, emergencies, and disasters. Implementing BC requires three IT measures: backup and replication of business state data, redundancy and failover of processing capabilities, and redundancy and failover of external interfaces.

Business Continuity Management (BCM) is an integrated management process that identifies threats, assesses their impact, and provides a framework for self‑recovery, protecting stakeholders, reputation, and value creation. BCM follows the PDCA (Plan‑Do‑Check‑Act) cycle.

The technical means to achieve BC are typically High Availability (HA) and Disaster Recovery (DR). HA aims to minimize downtime caused by planned maintenance or unexpected failures, often using cluster systems that combine redundant hardware and software components. Modern HA also incorporates Continuous Data Protection (CDP) to address data or logical faults.

DR involves rebuilding the IT service platform in a different location after a service interruption, covering both local recovery and remote reconstruction. HA handles single‑component failures within a data center, while DR addresses large‑scale failures requiring failover between data centers. HA operates at LAN scale; DR operates at WAN scale.

Tencent Cloud HA Solutions

Typical web system architecture layers are illustrated (image). Various HA structures are presented:

Structure 1: CLB + CVM in different Availability Zones (AZ) – By binding CVM instances from multiple AZs to a Cloud Load Balancer (CLB) within the same VPC, the service remains robust against a single AZ failure.

Structure 2: CLB + CVM + MySQL (same AZ) – Each CDB for MySQL instance provides a real‑time hot‑standby for high availability; a disaster‑recovery instance can be activated within seconds if the primary fails.

Structure 3: CLB + CVM + MySQL (different AZs) – Supports multi‑AZ deployment for MySQL (currently only in the Shenzhen Financial Zone).

CDB for MySQL Disaster‑Recovery Options

1. Ordinary instance architecture within the same AZ.

2. Cross‑city disaster‑recovery instance architecture (different AZs).

3. Cross‑region disaster‑recovery instance architecture.

These options enable cost‑effective improvement of service continuity and data reliability.

Cross‑AZ Disaster Recovery

Ensures that a CLB’s backend CVM instances are distributed across multiple AZs, preventing service loss when one AZ becomes unavailable. Users may purchase two CLB instances and use DNS to route traffic between them.

Cross‑Region Disaster Recovery

Addresses scenarios where an entire region fails. It involves DNS‑based failover (e.g., Cloud DNS + DNSPod) and may require inter‑region VPC peering for data synchronization.

Additional Data Synchronization Capabilities

• CVM Image – supports cross‑region replication (requires shutdown to create image). • COS Replication – future support for cross‑zone replication. • DTS (Data Transmission Service) – creates database sync tasks. • CLB Multi‑AZ – currently only in the Financial Zone. • MySQL Multi‑AZ – currently only in the Shenzhen Financial Zone.

The article concludes with reference diagrams for high‑availability architectures, including two‑data‑center, three‑center, and cross‑region designs.