What is a function of a spine-and-leaf architecture?

A spine-and-leaf architecture provides predictable latency because all devices are a consistent number of hops away from each other. Traffic from any leaf switch to any other leaf switch will always traverse one spine switch. This two-tier design (leaf to spine to leaf) ensures a uniform path length, which translates to predictable and often lower latency compared to traditional three-tier hierarchical designs that can have varying path lengths. Adding a layer of leaf switches doesn't inherently mitigate oversubscription; proper design and port capacity planning do. Spine-and-leaf architectures route all types of traffic, not exclusively multicast. The architecture does not limit the payload size of traffic within the leaf layer; that's a function of network protocols and device capabilities.