Multi-Site Azure Local Architectures

---

Overview

Deploy Azure Local clusters across multiple physical sites with synchronized operations, failover capabilities, and coordinated governance.

📝 Key Distinction

Azure Local does NOT support stretch clusters (single cluster spanning multiple sites). This page covers multi-cluster architectures where each site has its own Azure Local cluster, connected through replication and Arc management.

For rack-level high availability within a single cluster, see Rack-Aware Clustering.

Multi-Site Replication Topology

Figure 1: Azure Local multi-site deployment with synchronous and asynchronous replication across datacenters

---

Multi-Site vs. Rack-Aware Clustering

Feature	Rack-Aware Clustering	Multi-Site Architecture
Cluster Count	Single cluster	Multiple clusters
Physical Scope	2 racks (same campus, ≤1ms latency)	Multiple sites (any distance)
Storage	Single pool, synchronous	Separate pools, replicated
Replication	Built-in (Storage Spaces Direct)	Azure Site Recovery or Storage Replica
Failover	Automatic (zone-aware)	Manual or ASR-automated
Use Case	Room/building-level HA	Geographic DR
Status	Preview	GA

---

Multi-Site Topology Patterns

graph TB
    subgraph Pattern1[Hub-and-Spoke]
        Hub[Central Hub Site<br/>Control Plane]
        Spoke1[Spoke Site 1<br/>Workloads]
        Spoke2[Spoke Site 2<br/>Workloads]
        Spoke3[Spoke Site 3<br/>Workloads]

        Hub --> Spoke1
        Hub --> Spoke2
        Hub --> Spoke3
    end

    subgraph Pattern2[Peer-to-Peer]
        Site1[Site 1]
        Site2[Site 2]
        Site3[Site 3]

        Site1 <--> Site2
        Site2 <--> Site3
        Site3 <--> Site1
    end

    subgraph Pattern3[Tiered]
        Primary[Primary Site<br/>Control & Data]
        Secondary1[Secondary Site 1<br/>Replicas]
        Secondary2[Secondary Site 2<br/>Replicas]
        Tertiary[Tertiary Site<br/>DR & Archive]

        Primary --> Secondary1
        Primary --> Secondary2
        Secondary1 --> Tertiary
        Secondary2 --> Tertiary
    end

    style Pattern1 fill:#E8F4FD,stroke:#0078D4,stroke-width:2px,color:#000
    style Pattern2 fill:#FFF4E6,stroke:#FF8C00,stroke-width:2px,color:#000
    style Pattern3 fill:#F3E8FF,stroke:#7B3FF2,stroke-width:2px,color:#000

Distributed Hub-and-Spoke

• Central hub site (primary control plane)
• Spoke sites (application workloads)
• Replicated management
• Site-specific data residency

Peer-to-Peer Federation

• No central hub
• Direct site-to-site communication
• Distributed quorum
• Equal governance rights

Tiered Architecture

• Primary site (control plane & data)
• Secondary sites (read replicas, compute)
• Tertiary sites (DR & archival)
• Cascading replication

---

Multi-Site Architecture Comparison

---

Synchronization Mechanisms

Management Synchronization

• Cluster configuration sync
• Policy and governance distribution
• Certificate management coordination
• Update and patch orchestration

Data Replication

• Application data sync
• Database replication
• Storage synchronization
• Consistency requirements

Workload Distribution

• VM placement policies
• Traffic routing across sites
• Load balancing strategies
• Site affinity rules

---

Failover and Recovery

• Site failure scenarios

  • Single site outage
  • Network partition
  • Complete data center failure

• Recovery procedures

  • Failover automation
  • Manual intervention points
  • Data consistency verification
  • Service restoration order

---

Operational Considerations

• Monitoring across sites
• Log aggregation and correlation
• Remote support coordination
• Maintenance scheduling
• Update deployment sequence

---

See also: Connected Mode Architecture | Networking Deep Dive