Running database servers in the cloud presents several challenges, one of which is ensuring performance . A paper comparing the performance of SQL Server in its traditional and cloud versions (the latter known as SQL Database) highlights the need to adjust apps to suit the different approaches.
Database picture from Shutterstock
The cost benefits of moving to the cloud can be considerable, but that doesn't mean there aren't difficulties. In particular, you need to allow for the importance of recoding apps to match the performance characteristics of cloud systems:
The use of commodity hardware and servicing model in SQL Database implies that you should build commands to expect that some operations will fail and need to be retried. This requires thinking about how to split up operations into smaller chunks, avoiding long-lived temporary state, and making operations easily restartable.
Hit the link to read the full paper and further suggestions on how to tune database servers for each platform. The big lesson? As with any server migration, major planning is required.