How does a spine-and-leaf design scale when additional access ports are required?

In a spine-and-leaf architecture, adding a leaf switch is the primary method for scaling access ports. Each new leaf switch connects to every spine switch, maintaining the full mesh connectivity between the leaf and spine layers. This ensures consistent latency and bandwidth for all connected devices, as any leaf can reach any other leaf through a single spine hop. Adding a spine switch (option B) is for increasing overall fabric bandwidth and east-west traffic capacity, not directly for access ports. Adding a spine and leaf together (option A) isn't the most efficient way to scale just access ports. A single connection to a "core spine switch" (option D) is incorrect; there's no single "core" spine, and each leaf must connect to all spines for redundancy and optimal performance.