Skip to main content

Object Storage

CDB Object Storage stores any type of data — media files, logs, backups, AI/ML datasets — and serves it via the S3 API or SFTP protocol. Storage is organized into buckets with configurable access control and lifecycle management.

Storage types

Three storage types cover different performance and cost requirements. The location name when creating a storage indicates which type is provisioned:

  • S3 Fast — high-performance S3 storage optimized for AI/ML workloads, IoT telemetry, and latency-sensitive applications.
  • S3 Standard — general-purpose S3 storage for backups and content delivery.
  • SFTP — file-based storage accessed via SFTP, priced the same as S3 Standard.

Bucket and file management

S3 buckets hold files and folders, support direct HTTP access via URL, and work with any standard S3 client. Bucket contents are managed through the CDB Technical Web Portal or via AWS CLI and S3cmd.

Access control

Bucket access is controlled at two levels: ACLs set per-object public or private visibility, and bucket policies define rules by operation, IP address, or principal. Both are configured using AWS CLI or S3cmd.

Presigned URLs and lifecycle policies

Storage owners generate presigned URLs — time-limited download links for private files, valid for up to 7 days. Lifecycle policies automatically delete objects after a configured number of days, reducing storage costs for temporary data.

CDN and direct file access

S3 storage can serve as an origin for Orange's CDN resources, enabling cached content delivery to end users. Files are also accessible directly via a URL using the storage hostname and bucket path.