Comment 5 for bug 2024385

Thanks for the pointer. From my understanding, the agent in networking-baremetal is not used for switch configuration. The only purpose of this agent is to feed physnet mappings back to Neutron.

The recent switch configuration code in networking-baremetal is similar to what currently exists in NGS: it is synchronous and it runs in the context of the neutron/neutron-api process.

To be honest, I'm a bit surprised to see this netconf-based switch configuration code in networking-baremetal. What was the reasoning behind it? It seems to me that it would have been easier to integrate it in NGS, having a single "switch configuration" project with different backend drivers. That way, it would benefit from the existing infrastructure (config parsing, optional locking to control parallelism, performance optimisation, well-tested ML2 integration...)