Cloud Replicator Documentation
Index

Destinations Overview

Cloud Replicator supports four Azure cloud destinations, each designed for different use cases. This guide helps you understand when to use each destination and their key characteristics.

Supported Destinations

Destination API Type Best For
Azure Table Storage Table/Key-Value Cost-effective storage, simple queries
Azure Cosmos DB (NoSQL) Document Flexible schema, global distribution
Azure Cosmos DB (Table) Table Premium Table Storage features
Azure AI Search Search Index Full-text search, faceted navigation

Choosing the Right Destination

Azure Table Storage

Choose when you need:

  • Cost-effective storage for large amounts of data
  • Simple key-based queries (PartitionKey + RowKey)
  • Integration with Azure Functions or other Azure services
  • Pay-per-use pricing model

Characteristics:

  • Highly scalable NoSQL key-value store
  • Automatic scaling with no capacity planning
  • Low cost per GB stored
  • Limited query capabilities (primarily key-based)

Typical Use Cases:

  • Historical data archival
  • Simple reporting datasets
  • IoT or telemetry data storage
  • Backup and compliance archives

Azure Cosmos DB (NoSQL)

Choose when you need:

  • Global distribution with multi-region writes
  • Flexible document-based storage
  • Rich SQL-like query capabilities
  • Multiple consistency level options

Characteristics:

  • Fully managed NoSQL database
  • Guaranteed single-digit millisecond latency
  • 99.999% availability SLA
  • Automatic and instant scalability

Typical Use Cases:

  • Global applications requiring low latency
  • Real-time analytics dashboards
  • Event-driven architectures
  • Mobile and web application backends

Azure Cosmos DB (Table API)

Choose when you need:

  • Azure Table Storage compatibility with premium features
  • Existing applications using Table Storage SDK
  • Higher performance than standard Table Storage
  • Global distribution with Table API

Characteristics:

  • Table API compatible with Azure Table Storage
  • Premium throughput and latency
  • Turnkey global distribution
  • SLA-backed availability and performance

Typical Use Cases:

  • Migrating from Azure Table Storage
  • Applications requiring Table Storage API with better performance
  • Global deployments needing local data access

Choose when you need:

  • Full-text search across Business Central data
  • Faceted navigation and filtering
  • AI-powered search enhancements
  • Search-as-you-type functionality

Characteristics:

  • Managed search service
  • Powerful query syntax
  • Faceting, filtering, and scoring
  • Integration with Azure AI services

Typical Use Cases:

  • Product catalog search
  • Customer lookup applications
  • Document search portals
  • Knowledge base search

Feature Comparison

Feature Table Storage Cosmos DB NoSQL Cosmos DB Table AI Search
Query Type Key-based SQL-like Key-based Full-text
Per Day Series Yes Yes Yes No
SystemId as Key Optional Optional Optional Required
Deletion Tracking Yes Yes Yes No
Global Distribution Limited Yes Yes Limited
Cost Model Pay-per-use Provisioned RU/s Provisioned RU/s Tier-based

Data Structure by Destination

Table Storage / Cosmos DB Table

Data is stored using a PartitionKey + RowKey structure:

  • PartitionKey: Company name (cleaned) + optional postfix
  • RowKey: Primary key fields or SystemId

Example structure:

PartitionKey: "CRONUS_International"
RowKey: "10000"  (Customer No.)
Fields: Name, Address, City, etc.

Cosmos DB NoSQL

Data is stored as JSON documents in collections:

  • Collection: One per Business Central table
  • Partition Key: Company name field
  • Document ID: Primary key or SystemId

Example structure:

{
  "id": "10000",
  "__Company": "CRONUS_International",
  "Name": "Customer Name",
  "Address": "123 Main St"
}

Data is stored in search indexes:

  • Index: One per Business Central table
  • Document Key: SystemId (required)
  • Fields: Searchable, filterable, facetable based on type

Example structure:

{
  "SystemId": "abc-123-def-456",
  "Name": "Customer Name",
  "Address": "123 Main St"
}

Note: Unlike the other destinations, the company-name field is not added to Azure AI Search documents.

Mixing Destinations

You can replicate different tables to different destinations by overriding the default destination in each table mapping. Common patterns include:

Analytics + Search Pattern

  • Replicate transactional data to Cosmos DB for analytics
  • Replicate master data to AI Search for lookup functionality

Tiered Storage Pattern

  • Replicate frequently accessed data to Cosmos DB
  • Replicate historical data to Table Storage for cost efficiency

Multi-Purpose Pattern

  • Replicate the same table to multiple destinations for different use cases
  • Use the Index field in table mapping to create multiple mappings per table

Performance Considerations

Destination Write Speed Query Speed Scaling
Table Storage High Key-based fast, scans slow Automatic
Cosmos DB NoSQL High (depends on RU/s) Fast Manual/Auto
Cosmos DB Table High (depends on RU/s) Key-based fast Manual/Auto
AI Search Moderate Fast for search Tier-based

Cost Considerations

Azure Table Storage

  • Most cost-effective for storage
  • Pay per GB stored and transactions
  • No minimum cost

Azure Cosmos DB

  • Provisioned throughput (RU/s) model
  • Higher cost but better performance guarantees
  • Consider serverless for variable workloads
  • Tier-based pricing
  • Cost depends on document count and index size
  • Free tier available for small datasets