Elastic Compute Service:ECS data backup and high-availability architecture

Enhance cloud business resilience

Data durability protects recoverability through backup against logical errors or physical damage. Business continuity eliminates single points of failure through redundant architectures. Choose a strategy based on your business stage and disaster recovery requirements:

• Cost-sensitive — daily data protection needed.

  Build cost-effective backup without architectural changes. See Data backup and recovery.

• Growth stage — resilience against data center failures.

  Deploy multi-zone high availability so traffic automatically routes to healthy zones when one zone fails.

• City-level disaster resilience required.

  For industries such as financial services, gaming, and cross-border e-commerce, build a cross-region high-availability architecture to withstand regional disasters such as natural catastrophes or large-scale network outages.

Data backup and recovery

Recover from data loss caused by corruption, accidental deletions, or infrastructure failures.

1. Combine protection mechanisms based on features and protection scope.

   • Snapshots: Back up cloud disk data without client installation.

     Billed by snapshot type and capacity. See Snapshot billing.

     • Create snapshots: Create snapshots periodically and roll back to recover from accidental deletions.

       Or create a custom image, then restore by replacing the OS.

     • Replicate snapshots: Create an automatic snapshot policy with cross-region replication to handle region-level failures. On failure, create a data disk from the replicated snapshot and mount it to a standby instance.

   • ECS File Backup Basic Edition: File-level backup via an in-region client. Quickly restore deleted files.

     Each Alibaba Cloud account (including RAM users) shares a 100 GiB free quota across all regions. Excess usage is billed by total attached block storage capacity. See file-backup-essential-edition-benefit-description.

   • Cloud disks: Cross-zone and cross-region data backup with native cloud disk capabilities.

     • Regional ESSDs: Data replicates across zones within the region. On a zone-level failure, force-attach the disk to an instance in another zone.

       Billed by disk capacity. See Block storage devices.

     • Async replication: Replicates disk data to another disk in a different zone or region. On failure, trigger a manual failover, then attach the secondary disk to a standby instance.

       Includes capacity charges for the target disk. Cross-region replication incurs additional disk replication charges.

2. Set Recovery Point Objective (RPO) and Recovery Time Objective (RTO) targets. Quantify downtime and data loss costs with stakeholders, then validate targets through business continuity drills.

   RPO: Maximum tolerable data loss, measured in time.

   RTO: Maximum time from failure to full recovery.

   Important

   RPO and RTO are business metrics, not technical guarantees. Estimate your actual end-to-end values. Tighter targets cost more.

Multi-zone high availability

Single-instance deployment risks availability — any failure (such as hardware issues or process crashes) causes service interruption. Deploy multiple instances across zones with ALB for automatic failover via health checks.

• Application Load Balancer (ALB): Distributes traffic across healthy instances. Health checks remove unhealthy instances from rotation. ALB works with Auto Scaling (ESS) to replace failures and bring new instances online.

• Relational Database Service (RDS): High-availability editions use primary-standby architecture across zones to persist data.

Optimize performance and costs:

• Store static files such as images and scripts in Object Storage Service (OSS) and accelerate delivery with CDN.

• Use Auto Scaling (ESS) to handle traffic fluctuations. ESS adjusts capacity on demand and replaces failed instances automatically.

Cross-region high availability

Single-region deployment risks complete interruption during extreme disasters or large-scale outages. Build a cross-region high-availability architecture by extending cross-zone HA across regions with redundant systems, global traffic management, and data sync for automatic regional failover.

• Global Traffic Manager (GTM): Routes users by geography or latency. Health checks monitor each region; on failure, GTM redirects traffic to healthy regions via DNS.

• Data Transmission Service (DTS): Real-time bidirectional data sync across regions for active-active and disaster recovery scenarios.

Optimize performance and costs:

• Store static files such as images and scripts in Object Storage Service (OSS) and accelerate delivery with CDN.

• Use Auto Scaling (ESS) to handle traffic fluctuations. ESS adjusts capacity on demand and replaces failed instances automatically.

Business continuity drills

Drill regularly after your solution is in place:

1. Simulate failures. Run full recovery. Measure actual RPO and RTO against targets.

2. Verify that quotas, network configurations, and security policies in the failover zone or region are ready.

3. After recovery, verify data completeness, consistency, and application functionality.