Total spine capacity, bisection bandwidth and per-leaf share for a data centre fabric.
Every spine port terminates on a leaf, so the spine layer's aggregate port capacity is the ceiling on inter-leaf traffic. Bisection bandwidth halves that figure because traffic crossing an imaginary line through the fabric uses capacity on both sides. Bisection bandwidth is the number HPC and AI workloads are benchmarked against, and per-leaf capacity tells you whether adding leaves will starve the existing ones.
Fabric Bandwidth
spine capacity = spines × fabric ports per spine × port speed; bisection bandwidth is half of that.
spine capacity = spines × fabric ports per spine × port speed; bisection bandwidth is half of that. Every spine port terminates on a leaf, so the spine layer's aggregate port capacity is the ceiling on inter-leaf traffic. Bisection bandwidth halves that figure because traffic crossing an imaginary line through the fabric uses capacity on both sides.
Bisection bandwidth is the number HPC and AI workloads are benchmarked against, and per-leaf capacity tells you whether adding leaves will starve the existing ones.
This calculator takes 5 inputs: Spine switches, Fabric ports per spine, Fabric port speed, Leaf switches attached, Usable utilisation target. The pre-filled defaults are a realistic starting point — replace them with figures from your own environment for a result you can act on.