OpenStack Accelerator (Cyborg)

Cyborg is a general management framework for accelerators

Documentation for Operators

The documentation in this section is aimed at Cloud Operators needing to install or configure Cyborg.

For End Users

As an end user of Cyborg, you’ll use Cyborg to create and manage accelerators with either tools or the API directly.

Tools for using Cyborg

Information on the commands available through Cyborg’s Command Line Interface (CLI) can be found in this section of documentation.

Using the API

All end user (and some administrative) features of Cyborg are exposed via a REST API, which can be used to build more complicated logic or automation with Cyborg. This can be consumed directly, or via various SDKs. The Cyborg API has experienced an evolution from V1 API to V2 API. Cyborg introduced and landed a totally new DB modeling schema. for tracking cyborg resources in Stein release. The legacy v1 api does not match the new data model, which we changed pretty much. So cyborg introduced version 2.0 APIs and deprecated V1 APIs in Train release. Then in the Ussuri release, V1 APIs are removed, and full V2 APIs and microversion are supported. The following resource may help you get started with V2 APIs directly.

  • Cyborg API Reference: The complete reference for the accelerator API, including all methods and request / response parameters and their meaning.

Indices and tables