archivelogmode

Oracle has Concept material to help if you need this in-depth.  High level, data modification statements (DML) in Oracle are copied into a redo log.  When the statement is committed, the change is written into the physical datafile, and copied into one or more archive logs -- presuming the database instance is running in archivelog mode.

As slight points out, the redo mechanism is round-robin.  There is no long-term retention, except for the archive logs.

Also, by allowing multiple archivelog destinations, including to other hosts, Oracle provides redundancy.  Loss from a single backup medium is no longer catastrophic.

Loss from a single backup medium is no longer catastrophic.

As long as your Oracle DB runs with archivelog mode = on ;-)
This is/should be the default setting within productive environments. TEST, DEV and most other DB instances can be used without archivelog mode turned on.
At least this is how I/we used to work ;-)

Tony, we may have strayed from your question.  The "point" of archivelog mode is to enable the recovery of one's database to a system change event (SCN) or point in time. All data manipulation is recorded,  then replicated to provide multiple backups.  Archivelog mode is required for flashback.  Example, a user mistakenly drops a table.  With flashback enabled, the restoration of that table is quick and simple.

Many decision support / data warehouse systems, on the other hand, may not need to recover at will.  They might restore to a full backup point, then apply incremental changes -- or to re-run the extract/data/load (ETL) jobs that feed into the database.  It may be more practical to repeat a data import than to archive everything.

Does that help?

Thank you...