Slide 15
Slide 15 text
Ideas for when to use which architecture pattern
1 Cosmos DB Change Feed with Data Factory / Synapse Mapping Data Flow
• Prefer Low-code / No-code style
• Don’t want to / can’t use Synapse Link for Cosmos DB
• Need to process data with shorter latency than Synapse Link for Cosmos DB
2 Near-real time analytics with Synapse Serverless SQL pool & Power BI
• Appropriate for near-real time analytics
• BI with fully serverless, cost-effective way
3 Code based data processing with Synapse Spark pool
• Good to use with #2 & Custom partitioning (Preview)
• Large volume of data, which should be aggregated for analytics & BI
4 Code free data processing with Synapse Mapping Data Flow
• Good to use with #2 & Custom partitioning (Preview)
• Unlike #3, if Low-code / No-code style is preferable, use #4
5 Cost-effective archival of historical data
• Need to retain read-only copy cost-effectively