Azure Arc Data Services

What are Arc Data Services?

Azure Arc-enabled Data Services bring Azure database services to any infrastructure - on-premises, edge, or any public cloud. Get a managed database experience with evergreen updates, elastic scale, and Azure-consistent management while keeping data on-premises.

Available Services:

Azure SQL Managed Instance: Fully compatible SQL Server with PaaS benefits
PostgreSQL Hyperscale: Scalable PostgreSQL with Citus extension

← Back to Azure Arc Introduction

Available Services

Azure SQL Managed Instance

What It Is:

SQL Server compatibility (99%+)
Managed instance with automatic backups
Built-in high availability
Evergreen (always up-to-date)
Running on your infrastructure

Key Features:

T-SQL compatibility
Linked servers
CLR, SQL Agent, Database Mail
Full-text search
Elastic scale (Enterprise tier)
Point-in-time restore

Licensing:

Pay-as-you-go (Azure meter)
Bring-your-own-license (BYOL)
Azure Hybrid Benefit

PostgreSQL Hyperscale

What It Is:

PostgreSQL with Citus extension
Horizontal scaling (sharding)
Distributed tables and queries
Managed service experience

Key Features:

Scale out reads and writes
Parallel query execution
Real-time analytics
Multi-tenant applications
High throughput ingestion

Use Cases:

Multi-tenant SaaS applications
Real-time analytics dashboards
Time-series data
High-throughput OLTP

Deployment Modes

Directly Connected Mode

Characteristics:

Continuous or regular connection to Azure
Billing and usage data uploaded to Azure
Azure portal management available
Automatic updates

Requirements:

Outbound HTTPS to Azure
Azure subscription
Upload to Azure every 24 hours

Benefits:

Simplest to manage
Full Azure portal experience
Automated billing
Latest features first

Indirectly Connected Mode (Retired)

Historical Context:

This mode previously supported air-gapped environments with manual data export/import. Organizations requiring disconnected database services should now evaluate:

Azure Local Disconnected Operations — Local Azure portal and management
Traditional SQL Server/PostgreSQL — Self-managed databases on disconnected infrastructure

Managed Database Experience Anywhere

What “Managed” Means

Automated Operations:

Backups and restore
High availability and failover
Patch management and updates
Performance monitoring
Resource scaling

Reduced Admin Overhead:

No OS patching
No infrastructure management
Simplified disaster recovery
Built-in monitoring

Elastic Scale and High Availability

Scale Up/Down:

Adjust CPU and memory dynamically
No downtime for scaling
Pay only for what you use

Scale Out (Hyperscale):

Add read replicas
Shard data across nodes
Parallel query execution

High Availability:

Always-On Availability Groups (SQL MI)
Synchronous replication
Automatic failover (< 30 seconds)
Built-in health monitoring

Billing Model

Pay-As-You-Go

How It Works:

Billed based on vCore-hours consumed
Separate pricing for General Purpose and Business Critical tiers
Storage charged separately

Pricing Example (SQL MI):

General Purpose: ~$0.40/vCore/hour
Business Critical: ~$1.00/vCore/hour
Storage: ~$0.11/GB/month

Bring Your Own License (BYOL)

Requirements:

Active SQL Server license with Software Assurance
Azure Hybrid Benefit enrollment

Savings:

Up to 55% savings vs. pay-as-you-go
License mobility (move between environments)

Combined with Azure Hybrid Benefit:

Use existing on-premises licenses
Add free 180 days of Software Assurance
Further cost reduction

Migration from Traditional Databases

Assessment

Tools:

Azure Migrate for database assessment
Data Migration Assistant (DMA)
Azure SQL Migration extension

Assessment Output:

Compatibility issues
Feature parity analysis
Performance baseline
Sizing recommendations

Migration Methods

1. Backup and Restore:

For offline migrations
Full/diff/log backups
Point-in-time recovery
Minimal downtime: Hours

2. Log Shipping:

For near-zero downtime migrations
Continuous log replay
Manual cutover
Minimal downtime: Minutes

3. Distributed Availability Group (SQL MI):

Replicate from on-premises SQL Server
Bidirectional replication
Minimal downtime migration
Rollback capability

4. Azure Database Migration Service:

Online migrations
Minimal downtime
Automated process
Change data capture (CDC)

Post-Migration Validation

Performance:

Benchmark queries
Compare execution plans
Validate throughput

Functionality:

Test all application features
Validate integrations
Confirm backup/restore

Compliance:

Security controls
Audit logging
Encryption validation

Security and Encryption

Encryption at Rest:

Transparent Data Encryption (TDE) enabled by default
AES-256 encryption
Customer-managed keys (CMK) option

Encryption in Transit:

TLS 1.2+ for all connections
Force encryption option
Certificate validation

Access Control:

Azure AD authentication
SQL authentication
Role-based access control (RBAC)
Row-level security (RLS)
Dynamic data masking (DDM)

Auditing:

SQL audit to local storage or Azure
Threat detection
Vulnerability assessment
Compliance reporting (PCI-DSS, HIPAA, etc.)

Data Residency Guarantees

Data Stays On-Premises:

Database files remain on your infrastructure
Backups stored locally (or where you choose)
Query processing happens locally
No automatic replication to cloud

Metadata Sent to Azure (Direct Mode):

Billing and usage data
Performance metrics (aggregate)
Diagnostic logs (optional)
Configuration metadata

Compliance:

Meets GDPR requirements
HIPAA compliant
PCI-DSS ready
FedRAMP compatible

Audit Trail:

All data movement logged
Compliance reports available
Demonstrates data residency

Use Case Scenarios

Scenario 1: Financial Services Database Modernization

Challenge: 100+ SQL Server 2012 databases end of support, but data must stay on-premises.

Solution:

Deploy Arc SQL Managed Instance
Migrate databases using DMA
Implement Always-On AG for HA
Enable TDE and auditing

Results:

Modernized databases without cloud migration
Pay-as-you-go reduced costs by 40%
99.95% availability SLA
Passed compliance audits
Reduced admin overhead by 50%

Scenario 2: Healthcare SaaS Multi-Tenant Application

Challenge: SaaS app needs to scale to 1000s of tenants with high performance.

Solution:

Deploy PostgreSQL Hyperscale
Shard by tenant ID
Scale to 10+ worker nodes
Parallel query execution

Results:

10x query performance improvement
Seamless scaling to 5000+ tenants
Sub-second query response times
Reduced infrastructure costs (better utilization)

Scenario 3: Retail Point-of-Sale System

Challenge: Chain of 500 stores needs local database at each store with central reporting.

Solution:

Deploy SQL MI at each store location
Local transactions (no cloud dependency)
Nightly replication to central data warehouse
Azure portal for monitoring all 500 instances

Results:

Zero downtime from internet outages
Real-time POS transactions
Centralized reporting and analytics
Managed from single pane of glass

Best Practices

1. Right-Size from the Start:

Use assessment tools
Start with actual workload metrics
Monitor and adjust

2. Implement HA from Day 1:

Always-On AG for SQL MI
Multiple coordinator nodes for Hyperscale
Test failover procedures

3. Plan for Growth:

Size for 2-3 year growth
Consider scaling options
Budget for storage growth

4. Backup Strategy:

Define RPO and RTO
Test restores regularly
Implement offsite backups

5. Monitor Proactively:

Set up alerts
Review performance metrics weekly
Capacity planning

6. Leverage Azure Hybrid Benefit:

Use existing SQL licenses
Significant cost savings
License mobility

Cost Considerations

Total Cost of Ownership:

Arc Data Services:

Lower than Azure SQL Database for sustained workloads
Lower than self-managed on-premises (reduced admin)
Flexible with BYOL

Example Monthly Cost (SQL MI, 16 vCores):

Pay-as-you-go:

General Purpose: ~$4,600/month
Business Critical: ~$11,500/month

With Azure Hybrid Benefit:

General Purpose: ~$2,100/month
Business Critical: ~$5,200/month

Break-Even vs. Azure SQL Database:

Typically 6-12 months for sustained workloads
Faster for 24/7 production databases

Troubleshooting

Deployment Failures:

Verify Kubernetes resource availability
Check storage class configuration
Validate network connectivity
Review controller logs

Performance Issues:

Check resource allocation (CPU, memory)
Review query execution plans
Validate storage IOPS
Monitor network latency

Connectivity Problems:

Verify service endpoint configuration
Check firewall rules
Validate DNS resolution
Test from application server

Next Steps

External Resources:

Last Updated: October 2025