RabbitMQ-less OpenStack Cluster Supporting 14,000 Hypervisors in One Region

Explore a technical conference presentation that details the development and implementation of a revolutionary RPC mechanism designed to replace RabbitMQ in large-scale OpenStack clusters. Learn how LY Corporation successfully scaled their OpenStack infrastructure to support over 14,000 hypervisors in a single region by implementing a distributed software architecture. Discover the key differences in Day2 operations between traditional RabbitMQ and the new distributed system, along with valuable insights gained from two years of managing this extensive infrastructure. Gain practical knowledge about overcoming common bottlenecks in large-scale OpenStack deployments and understand the technical considerations behind eliminating RabbitMQ dependencies while maintaining robust cluster performance.