Backups
Object storage vs SFTP backup storage: choose by workflow
S3-compatible storage and SFTP storage can both hold a second copy of data, but they expose very different interfaces. The right destination depends less on which protocol sounds newer and more on what the backup or application already knows how to use safely.
SFTP looks like remote files
SFTP runs over SSH and exposes a filesystem-oriented workflow. Backup software opens paths, creates directories and transfers files much like it would to another server, while the provider can restrict the account to file transfer with no interactive shell.
That model is approachable for scripts and tools that already produce backup files or directory trees. It can also be easier to inspect manually during a restore because the destination resembles ordinary files.
S3 works with buckets and objects
Object storage uses bucket and object APIs over HTTP, not filesystem syscalls. Applications can write objects directly, attach metadata, use multipart transfers and create temporary presigned access without mounting a remote filesystem.
This is often a better fit for modern backup products, cloud-native applications and software that was designed around object semantics from the beginning.
Credential scope differs
An SFTP service commonly scopes an SSH key to one account and one filesystem root. An S3 service normally scopes access through project, bucket or policy boundaries and separate access/secret credentials.
Both can be secure when isolation is enforced by the backend. The dangerous design is shared credentials that rely on customers politely staying inside a naming prefix.
Restore behavior should drive the choice
If recovery means copying an archive back to one server, SFTP may be enough. If recovery software expects an S3 repository with object enumeration, multipart download or object version behavior, forcing it through SFTP can remove useful native features.
Run a restore test with the same software that creates the backups. Successful upload is not evidence that the selected protocol will support the recovery workflow under pressure.
Keep the products separate
A hosting provider should not relabel one storage pool as both SFTP and S3 merely for marketing. Capacity enforcement, credentials, abuse controls, billing reconciliation and failure modes are different enough to deserve separate service lifecycle logic.
DotMoose therefore treats SFTP Backup Storage and S3-compatible Object Storage as different products. A customer can use either or both based on the recovery design.