A lifecycle allows you to model the movement of data from expensive, rapid storage, through various forms of cheaper slower storage as the data ages and access requirements diminish. The period during which data remain in each kind of storage are modeled as phases, which are associated with tablespaces. The data movement itself is performed at the level of partitions.
The following diagram illustrates a lifecycle of five years, which is divided into three phases:
- Phase 1 (3 months) - high performance (tier-1) storage for new data that is frequently accessed.
- Phase 2 (9 months) - nearline (tier-2) storage for data from the last year.
- Phase 3 (48 months) - historical (tier-3) storage for data that is infrequently accessed but which must be retained.
The data is packaged in partitions, which each contain one month of data:
PowerDesigner can generate all the necessary scripts to automate all this data movement. In the example above, one script will be generated for every month of the lifecycle. At the point illustrated in the picture, the script will:
- Move partition P1 from the tablespace associated with Phase 1 to the tablespace associated with Phase 2
- Create a new partition, P4, to begin collecting new table rows in the tablespace associated with Phase 1
As the data ages, scripts will additionally treat the movement of data aged more than one year from the tablespace associated with Phase 2 to the tablespace associated with Phase 3.
Once a lifecycle is put in place, you can generate scripts to perform data movement indefinitely. Additional scripts are generated to regularly purge data that arrive at the end of their lifecycle.