Object storage

DT Cloud Services offers SwiftStack object storage (an OpenStack project) with a native and geo-redundant deployment. It is suitable for unstructured data without any known total data size limit, such as files, videos, analytics data, web content, backups, images and virtual machine snapshots.

Data is accessed via HTTPS using the SwiftStack API or Swift S3 API. Objects have a software layer that manages metadata, access permissions and intelligent behaviors directly at the data level.

At the core of SwiftStack is a cloud file system that runs on standard hardware, providing flexible and cost-effective storage solutions. It is highly scalable and optimized for durability and availability across the entire data set. The SwiftStack Controller manages all storage resources and data allocation policies, also across multiple geographical regions for data protection and optimized data access.

SwiftStack is designed for large scale storage with high availability and 12 nines of durability. Applications can access and consume storage using either SMB (Windows, stateful) or NFS (Linux, stateless) network protocols for file access. Cluster nodes are deployed in multiple geographic regions for data protection and optimized local access.

The S3 object storage API

The S3 REST API is emulated on top of the object storage with the following operations supported:

  • GET Service

  • DELETE Bucket

  • GET Bucket (List Objects)

  • PUT Bucket

  • DELETE Object

  • GET Object

  • HEAD Object

  • PUT Object

  • PUT Object (Copy)

With its high level of data portability, scalability, performance, and interoperability with machine-learning libraries, the Swift object storage is well suited for cloud-based AI solutions.

Current limitation: object storage is not yet reachable via internet but only through DT internal networks.

Additional Resources

SwiftStack API

SwiftStack S3 Support