•Data
warehouses containing large volumes of historical data
–Compress all
of the older data in a data warehouse
–Integrate
compression into the ‘rolling window’ paradigm
–For example, most recent 3 months of data
could be stored uncompressed and the previous 21 months could be stored
compressed
•Materialized
views and other derived data sets
•Generally,
compression should be applied to data that is infrequently
updated
•