Tune replication performance with AWS DMS for an Amazon Kinesis Data Streams target endpoint – Part 1

This article discusses how to tune the performance of AWS Database Migration Service (DMS) when replicating data to Amazon Kinesis Data Streams as the target. It covers the multi-threaded full load and change data capture (CDC) settings available in DMS and considerations for tuning related parameters for better performance.

Specifically, the article covers:

• Solution overview: Architecture for full load and CDC phases when using Kinesis Data Streams as the DMS target, including how DMS distributes data across Kinesis shards based on a partition key
• Considerations for parallel load and apply settings: Ordering of records, understanding source database workloads, determining number of Kinesis shards and partition key, and sizing the DMS replication instance
• Example use case: Demonstrating how to determine parallel apply settings based on source workload and number of Kinesis shards
• General best practices: Guidelines for choosing appropriate settings based on factors like ordering requirements, data distribution, and capacity needs
• Conclusion: Summary and preview of the next parts in the blog series