Maintaining Separate Database Connection Sources with Aurora Multi-Master
Customer with the following question regarding Aurora multi-primary.
Will an application still have to maintain separate database connection sources, like with the Aurora read replicas? My hope is that the multi-primary would provide just one endpoint, making it seamless to the application.
- 新しい順
- 投票が多い順
- コメントが多い順
Providing a single "seamless" endpoint to applications is an anti-pattern that will not be supported with Aurora Multi-Master.
Aurora Multi-Master is optimized for workloads where multiple writers can handle non-conflicting workloads - writes on either physically or logically separated data from multiple nodes so they don't produce conflicts, or minimize the risk thereof. This requires application awareness of the cluster topology and individual nodes, so the app can direct traffic to the various writers based on a workload specific non-conflicting segmentation of the data.
Aurora Multi-Master is not recommended for customers that desire a drop in replacement for a single master that would give them "seamless" multi-node write scaling by load balancing behind a single endpoint. This is a use case that is expected to have significantly high rates of conflicts, which will detrimentally impact performance.
関連するコンテンツ
- 質問済み 6年前
AWS公式更新しました 3年前
