Every operation in OpenStack requires validating a token against the Keystone server. To have a multisite deployment, either you have one Keystone database with multiple regions, or you have multiple Keystones. Both approaches limit the operator’s ability to manage multisite deployments. How can we make Keystone actively enable edge computing instead of being a stumbling block?
In this session, we will discuss what is possible now and how that limits the multisite deployments that providers wish to pursue. We will also address what is possible by rearchitecting Keystone to better support and managing multisite deployments.
How Keystone works now, what are the stumbling blocks for multisite, and how we plan to address them.