Data Archive uses the Data Vault Service to move inactive data to a Data Vault.
When you use Data Archive, you can archive data in multiple ways. You can archive data to partitions in the same database, to another database, or to a Data Vault. The Data Vault is an immutable, highly compressed file structure that you can query using standard SQL syntax.
The Data Vault provides the following features and benefits:
Storage efficiency, improved application performance, speed and ease of user access, and low overall cost of ownership.
The Data Vault stores data in a fraction of the space that would be necessary using a relational format, offering order-of-magnitude savings in storage costs.
Unlike tape backup archives, which require tape retrieval and loading as well as often complex and time-consuming searches, end users can run queries directly against the archive using the same Business Intelligence tools that they use for other analytical tasks.
Unlike other compression technologies that require full uncompression of data before searching, you can run queries against the compressed files in Data Vault, and the result set is uncompressed.
End users can access Data Vault directly, without the need to involve a DBA or archivist in the process.
Data can be loaded and compressed into archived table form from any relational or flat file data source. The Data Vault stores the compressed, immutable data in the form of SCT files. The compressed data can be extracted by an SQL routine from either an application script or through a query tool, loaded into any target data database, and joined there with existing tables if required.
The degree of data compression is a function of the type of data. Since tokenization is a key part of the process, the more duplication there is in the source data, the higher the compression ratio. This also implies that the greater the amount of data–and therefore the higher the probability of duplication–the higher the resulting compression ratio.
The Data Vault cannot be updated. However, once data has been retrieved from the Data Vault and loaded into another data structure, it can be manipulated in all the usual ways to perform operations such as "what-if" analyses.