Designed by Spectra Logic, DS3 is the first native REST-based interface to deep storage which enables the easy archival of large amounts of bulk data. It is a data transport and communication interface that allows software clients to direct and manage "bulk" storage read or write operations to and from deep storage. DS3 utilizes the standard S3 command set plus additional commands specifically designed to optimize the transport of data objects to and from tape. These additional commands enable modern architecture frameworks to easily and efficiently communicate with massively scalable deep storage technologies. The DS3 interface utilizes the traditional REST client/server architecture and high-level PUT and GET commands to move objects to and from deep storage. The interface is media agnostic such that over long data retention periods data can be migrated from older storage types to newer ones transparently.
DS3 offers a simple and modern approach to transferring large bulk files to and from deep storage. Spectra will be providing application specific clients and plug-ins to leverage the DS3 interface for bulk transfer of data objects to tape and other media. Developers can also develop additional clients or plug-ins using the DS3 superset of the S3 REST object storage API by leveraging the Spectra Logic Developer Program DS3 SDK documentation, simulators, code examples, FAQ's and other development tools.
DS3 enables you to efficiently move large quantities of data with minimal cost and effort. The DS3 communication interface allows the client to direct and manage "bulk" storage read or write operations of data objects. The first implementation supports bulk object storage operations with tape for accessibility to the lowest cost media option. To help ensure you can maintain your data in deep storage for indefinite periods of time, the interface provides high level PUT and GET commands for data objects and is agnostic to storage media types.
DS3 in combination with the Spectra BlackPearl™ Deep Storage Appliance provides a massively and infinitely scalable, reliable and low cost data object storage repository. Data objects are organized in S3-like buckets that represent a collection of related objects with common policy settings. Objects in buckets are completely independent of each other, and the deep storage system will dynamically allocate tapes for each bucket on an as-needed basis and stream to tape for optimized performance. These deep storage tape systems can start as small as 15TB and grow seamlessly to Exascale.
Have questions? Ready for next steps? Contact Us!