What are two design constraints in a standard spine and leaf architecture? (Choose two.)

Spine switches do not connect to each other - inter-spine links are prohibited because the spine tier functions exclusively as an interconnect layer for leaf switches.

Endpoints connect only to leaf switches, not to spine switches - spine switches carry only inter-leaf transit traffic.

Every spine switch must connect to every leaf switch to ensure that any server on a leaf can reach any other server in exactly two hops, maintaining predictable latency and equal-cost paths.

Leaf switches do not connect to each other - inter-leaf links are prohibited because all inter-leaf traffic must traverse the spine tier to maintain the two-hop topology guarantee.

Every leaf switch must connect to every spine switch for the same reason - to guarantee that east-west traffic between servers on different leaves always has a direct equal-cost multipath route through the spine tier.