Many businesses are eliminating the cost of keeping paper-based records by switching to web-based records.
However, not all of the data in a data warehouse is equally valuable to an organization. In general, the most recent data in a warehouse is much more likely than older data to be accessed by queries and maintenance processes or to be updated. Such data is therefore called hot. As time goes by, data tends to cool off, becoming warm and later cold , meaning that the probability that users access or update this data significantly decreases. The data must still be available, however, for regulatory requests, audits, and long-term research. Another important characteristic of requests for colder data is that users do not typically insist on optimal performance for these requests. Because strong performance for these queries is not essential, you can place colder data on slower, less expensive types of storage devices.
A warehouse can contain several different temperature tiers (hot, warm, cold, dormant). In general, the number of temperature tiers is tied to the number of different types of storage devices that are attached to the warehouse. For example, you might store hot data on new, solid-state drives (SSD); warm data on new, fast magnetic storage devices; and cold and dormant data on older, less efficient magnetic storage devices.
The definition of each data temperature depends on the specific environment, but data temperatures usually fall into fairly common categories. The following chart provides some guidelines for classifying data by temperature: