Repository-based backups are created by replicating data to a disk-based store. Restore point snapshots are created in the repository as required. The repository snapshot is assigned a retention period and is available for full or partial restore until it is retired ( deleted from the store). A Generation 1 repository store contents, created using a batch mover, can also be tiered to HCP for long term retention. For Generation 2 HCP, there is no need for a repository between the source node and HCP, backups can go straight from the source node to HCP. However, tiering is not supported on Generation 2 HCP.
The concept of Generation 1 and Generation 2 HCP nodes applies to Protector only and differentiates between an older HCP interface implementation and the current one. Generation 1 HCP is deprecated and should not be used for new configurations. In addition, you cannot upgrade Generation 1 implementations to Generation 2. There is difference between a backups to Generation 2 HCP and to HCP Cloudscale.
The concept of Generation 1 and Generation 2 repositories is also a Protector concept. Again. Generation 1 repositories are deprecated and should not be used for new configurations and cannot be upgraded to Generation 2 repositories. Generation 2 repositories have many advantages over Generation 1 repositories especially around parallelism.