One archive is created for each backup plan. Once you assign a new plan to a machine and start the first backup, a new archive is created.
The software uses an "always incremental" archive. Unlike older formats where incremental changes are written into individual files in the backup location, in an always incremental archive type there is physically one file in the archive location while the incremental changes are written to this file directly. This technology helps to avoid complications in retention/deletion of backups specific to the older formats where either a backup with dependencies on it from subsequent backups can be deleted only when all subsequent backups are due to deletion or a space-consuming consolidation process is required.
When a certain backup expires due to the pre-defined retention rules, the backup algorithm just marks these outdated backup blocks as “free” ones. The blocks of expired backups with dependencies (which are needed to restore the newer backups) are not marked as “free” to ensure the archive consistency. The always incremental archive total size includes the size of the “used” blocks and the size of the “free” blocks: the initial archive still occupies the same space on the storage as before, however all newer backups will be written to the “free” blocks first, and only if all the “free” blocks are filled, the total size of the archive will be increased.