Agree on the miss-traffic math — even a high HIT ratio still pins you to origin quality. What we see operating our own edges is that origin placement only pays off if the control plane can treat nodes as inventory: push sites, rotate certs, and purge without a ticket per PoP. We open-sourced an independent CDN control plane (CDNCA) for self-hosted nodes and a plugin marketplace; marketplace features roll out release by release. Did bare metal change your purge or failover story as much as latency?