When performance problems occur, one bottleneck is backing up the data on the hard drive. We can counter this by setting the saving policy. Depending on the selected setting, it is a trade-off between performance and security. With my suggestion I would like to defuse this tension.
With the SQL-Server from Microsoft it is possible to split tables and index on several hard disks. In this way, many more write and read operations can be executed simultaneously. This would make the backup faster during operation. In addition, the system is available faster after a restart. Depending on how this is implemented, backups could also be created and restored faster.
Functionality:
For each cube, a developer can decide whether to partition it. If yes, he chooses an entity for the partitioning. Depending on the available partitions, a script defines which entity elements are stored on the respective partition.
Example:
Our selected entity has 5 elements. These could be product areas. The product areas A, B and C will be written to partition 1. D and E are stored on the second partition.
If needed, more partitions can be added to scale the system.
