Skip to main content

Regional Availability

Alactic AGI can be deployed in multiple Azure regions worldwide. This guide helps you select the optimal region based on service availability, latency, compliance, and cost considerations.

Service Requirements by Region

Alactic AGI requires these Azure services in your selected region:

  • Azure OpenAI Service (GPT-4o or GPT-4o mini models)
  • Azure Cosmos DB (Serverless tier)
  • Azure Storage (Standard LRS)
  • Virtual Machines (D-series or B-series)
  • Azure Key Vault

Critical: Not all services are available in all regions.

Fully Supported Regions

These regions support all required services for Alactic AGI deployment:

Americas

East US (Virginia)

  • Status: Fully Supported
  • Azure OpenAI: GPT-4o, GPT-4o mini, Embeddings
  • Cosmos DB: Serverless available
  • VM Availability: Excellent (D, E, F series)
  • Latency from US East Coast: <5ms
  • Recommended For: US-based deployments, lowest latency for East Coast

East US 2 (Virginia)

  • Status: Fully Supported
  • Azure OpenAI: GPT-4o, GPT-4o mini, Embeddings
  • Cosmos DB: Serverless available
  • VM Availability: Excellent
  • Latency from US East Coast: <10ms
  • Recommended For: Redundancy with East US, US East Coast

West US (California)

  • Status: Fully Supported
  • Azure OpenAI: GPT-4o, GPT-4o mini, Embeddings
  • Cosmos DB: Serverless available
  • VM Availability: Good
  • Latency from US West Coast: <5ms
  • Recommended For: US West Coast deployments

West US 2 (Washington)

  • Status: Fully Supported
  • Azure OpenAI: GPT-4o, GPT-4o mini, Embeddings
  • Cosmos DB: Serverless available
  • VM Availability: Excellent
  • Latency from US West Coast: <10ms
  • Recommended For: Pacific Northwest, Alaska

West US 3 (Arizona)

  • Status: Fully Supported
  • Azure OpenAI: GPT-4o, GPT-4o mini, Embeddings
  • Cosmos DB: Serverless available
  • VM Availability: Good
  • Latency from US Southwest: <5ms
  • Recommended For: Southwest US, low-cost alternative

Canada Central (Toronto)

  • Status: Fully Supported
  • Azure OpenAI: GPT-4o, GPT-4o mini, Embeddings
  • Cosmos DB: Serverless available
  • VM Availability: Good
  • Latency from Canada: <5ms
  • Recommended For: Canadian data residency requirements

Brazil South (São Paulo)

  • Status: Fully Supported
  • Azure OpenAI: GPT-4o mini, Embeddings (GPT-4o limited)
  • Cosmos DB: Serverless available
  • VM Availability: Moderate
  • Latency from South America: <20ms
  • Recommended For: South American deployments, Brazilian data residency

Europe

West Europe (Netherlands)

  • Status: Fully Supported
  • Azure OpenAI: GPT-4o, GPT-4o mini, Embeddings
  • Cosmos DB: Serverless available
  • VM Availability: Excellent
  • Latency from Western Europe: <5ms
  • Recommended For: European deployments, GDPR compliance

North Europe (Ireland)

  • Status: Fully Supported
  • Azure OpenAI: GPT-4o, GPT-4o mini, Embeddings
  • Cosmos DB: Serverless available
  • VM Availability: Good
  • Latency from Northern Europe: <10ms
  • Recommended For: UK, Ireland, Northern Europe

UK South (London)

  • Status: Fully Supported
  • Azure OpenAI: GPT-4o, GPT-4o mini, Embeddings
  • Cosmos DB: Serverless available
  • VM Availability: Excellent
  • Latency from UK: <5ms
  • Recommended For: UK data residency, GDPR compliance

France Central (Paris)

  • Status: Fully Supported
  • Azure OpenAI: GPT-4o, GPT-4o mini, Embeddings
  • Cosmos DB: Serverless available
  • VM Availability: Good
  • Latency from France: <5ms
  • Recommended For: French data residency

Germany West Central (Frankfurt)

  • Status: Fully Supported
  • Azure OpenAI: GPT-4o, GPT-4o mini, Embeddings
  • Cosmos DB: Serverless available
  • VM Availability: Good
  • Latency from Germany: <5ms
  • Recommended For: German data residency

Sweden Central (Gävle)

  • Status: Fully Supported
  • Azure OpenAI: GPT-4o, GPT-4o mini, Embeddings
  • Cosmos DB: Serverless available
  • VM Availability: Good
  • Latency from Nordic countries: <10ms
  • Recommended For: Nordic data residency

Asia Pacific

East Asia (Hong Kong)

  • Status: Fully Supported
  • Azure OpenAI: GPT-4o, GPT-4o mini, Embeddings
  • Cosmos DB: Serverless available
  • VM Availability: Good
  • Latency from Hong Kong: <5ms
  • Recommended For: Hong Kong, Southern China

Southeast Asia (Singapore)

  • Status: Fully Supported
  • Azure OpenAI: GPT-4o, GPT-4o mini, Embeddings
  • Cosmos DB: Serverless available
  • VM Availability: Excellent
  • Latency from Southeast Asia: <10ms
  • Recommended For: ASEAN countries, Singapore data residency

Australia East (Sydney)

  • Status: Fully Supported
  • Azure OpenAI: GPT-4o, GPT-4o mini, Embeddings
  • Cosmos DB: Serverless available
  • VM Availability: Good
  • Latency from Australia: <5ms
  • Recommended For: Australia, New Zealand

Japan East (Tokyo)

  • Status: Fully Supported
  • Azure OpenAI: GPT-4o, GPT-4o mini, Embeddings
  • Cosmos DB: Serverless available
  • VM Availability: Excellent
  • Latency from Japan: <5ms
  • Recommended For: Japan data residency

Korea Central (Seoul)

  • Status: Fully Supported
  • Azure OpenAI: GPT-4o, GPT-4o mini, Embeddings
  • Cosmos DB: Serverless available
  • VM Availability: Good
  • Latency from Korea: <5ms
  • Recommended For: South Korea data residency

India Central (Pune)

  • Status: Fully Supported
  • Azure OpenAI: GPT-4o mini, Embeddings (GPT-4o limited)
  • Cosmos DB: Serverless available
  • VM Availability: Good
  • Latency from India: <10ms
  • Recommended For: India data residency

Middle East & Africa

UAE North (Dubai)

  • Status: Partial Support
  • Azure OpenAI: Limited availability (check current status)
  • Cosmos DB: Serverless available
  • VM Availability: Good
  • Latency from UAE: <5ms
  • Recommended For: UAE, Middle East (verify OpenAI availability)

South Africa North (Johannesburg)

  • Status: Partial Support
  • Azure OpenAI: Limited availability
  • Cosmos DB: Serverless available
  • VM Availability: Moderate
  • Latency from South Africa: <10ms
  • Recommended For: South Africa (verify OpenAI availability)

Regions with Limited Support

Azure OpenAI Not Available

These regions do not support Azure OpenAI Service:

  • Central US
  • South Central US
  • West Central US
  • North Central US
  • Norway East
  • Switzerland North
  • UAE Central

Workaround: Deploy in nearby supported region, use regional data replication.

Cosmos DB Serverless Not Available

These regions do not support Cosmos DB Serverless tier:

  • Gov regions (US Gov, China)
  • Some emerging regions

Workaround: Use provisioned throughput (higher cost) or deploy in supported region.

Region Selection Decision Factors

1. Latency Requirements

User Location vs. Region Latency:

User LocationRecommended RegionTypical Latency
US East CoastEast US, East US 2<10ms
US West CoastWest US, West US 2<10ms
CanadaCanada Central<10ms
UKUK South<10ms
Western EuropeWest Europe, France Central<15ms
GermanyGermany West Central<10ms
NordicSweden Central<15ms
SingaporeSoutheast Asia<10ms
AustraliaAustralia East<10ms
JapanJapan East<10ms
KoreaKorea Central<10ms
IndiaIndia Central<20ms
Hong KongEast Asia<10ms
BrazilBrazil South<20ms
Middle EastUAE North<20ms

Performance Impact:

  • <20ms: Excellent user experience
  • 20-50ms: Good user experience
  • 50-100ms: Acceptable for batch processing
  • >100ms: Noticeable delays, not recommended

2. Data Residency and Compliance

GDPR Compliance (EU Data Residency):

  • West Europe (Netherlands)
  • North Europe (Ireland)
  • UK South (London)
  • France Central (Paris)
  • Germany West Central (Frankfurt)
  • Sweden Central (Sweden)

Country-Specific Data Residency:

  • UK: UK South
  • France: France Central
  • Germany: Germany West Central
  • Canada: Canada Central
  • Australia: Australia East
  • Japan: Japan East
  • Korea: Korea Central
  • Brazil: Brazil South

3. Cost Variations

Azure pricing varies by region. Examples for Standard_D2s_v3 VM:

RegionVM Cost/MonthRelative Cost
East US$140Baseline
West US$140Same
West Europe$154+10%
UK South$161+15%
Australia East$182+30%
Japan East$175+25%
Brazil South$210+50%

Cost Optimization: Choose US East or West US for lowest infrastructure costs if compliance allows.

4. Service Availability and Quotas

Azure OpenAI Capacity: Some regions have limited Azure OpenAI capacity due to high demand:

  • Check current availability before deployment
  • Have backup region selected
  • Request quota increase if needed

VM Quota: Default quotas vary by region and subscription type:

  • Standard subscription: 10-20 vCPUs per region
  • Enterprise subscription: Higher quotas
  • Request increase via Azure Portal if needed

5. High Availability and Disaster Recovery

Multi-Region Deployment (Enterprise Only): Deploy in two regions for redundancy:

Primary-Secondary Pairs:

  • East US + West US (US coverage)
  • West Europe + North Europe (EU coverage)
  • Southeast Asia + East Asia (APAC coverage)

Active-Active: Traffic distributed across both regions
Active-Passive: Secondary region standby, failover on primary failure

How to Verify Region Support

Method 1: Azure Portal

  1. Go to Azure Portal
  2. Click "+ Create a resource"
  3. Search for "Azure OpenAI Service"
  4. Check "Region" dropdown for availability

Method 2: Azure CLI

# Check Azure OpenAI availability
az cognitiveservices account list-skus \
--kind OpenAI \
--location eastus

# Check Cosmos DB availability
az cosmosdb locations list \
--query "[?name=='East US']"

# Check VM availability
az vm list-skus \
--location eastus \
--size Standard_D \
--output table

Method 3: Azure Resource Explorer

  1. Visit resources.azure.com
  2. Navigate to Providers
  3. Check service availability by region

Regional Performance Benchmarks

Average response times for document processing:

RegionPDF ProcessingURL ScrapingModel Inference
East US2.1s1.8s0.4s
West US2.3s2.0s0.4s
West Europe2.4s2.1s0.5s
UK South2.2s1.9s0.5s
Southeast Asia2.8s2.4s0.6s
Australia East3.1s2.7s0.7s
Japan East2.6s2.3s0.5s

Note: Times measured from VM to Azure services within same region.

Changing Regions After Deployment

Important: You cannot change the region of an existing deployment. To move regions:

  1. Export Data

    • Backup Cosmos DB data
    • Download any stored documents
  2. Delete Current Deployment

    • Delete resource group
    • Or delete individual resources
  3. Redeploy in New Region

  4. Import Data

    • Restore Cosmos DB data
    • Upload documents to new storage

Downtime: Plan for 30-60 minutes of downtime during migration.

Best Practices

For Production Deployments:

  • Choose region closest to majority of users
  • Verify Azure OpenAI capacity before deployment
  • Consider data residency requirements
  • Plan for disaster recovery if critical
  • Monitor latency after deployment

For Development/Testing:

  • Use cheapest region (typically East US)
  • Data residency not critical
  • Can redeploy easily if needed

For Multi-Tenant SaaS:

  • Deploy in multiple regions
  • Route users to nearest deployment
  • Use Azure Front Door for global routing