HANA demands optimal distribution of data across all the HANA blades for best performance. A proper table distribution helps for more optimal load balancing and better parallelization. In this blog we will cover only the table distribution part. Table partitioning optimization will be described in another SCN article.
Here are several basic rules for general DB table distribution which this app follows:
Here are several basic rules for general DB table distribution which this app follows:
- Large tables should not be replicated.
- Small tables can exist on different nodes to leverage more optimal joins and prevent network transfer between the DB nodes.
- Tables are distributed as evenly as possible.
