Which data center topology gives each server a similar number of hops to any other?
Why D is the answer
Spine-leaf
Every leaf connects to every spine.
In detail
- Every leaf connects to every spine, so paths are even.
- It gives predictable paths.
Use case
An illustrative scenario, not a real incident.
Traffic between any two servers crosses a similar number of hops, because each leaf connects to each spine.
See it in different ways
flowchart TD S1["Spine"] --- L1["Leaf"] S1 --- L2["Leaf"] S2["Spine"] --- L1 S2 --- L2 L1 --- N1["GPU server"] L2 --- N2["GPU server"]
sequenceDiagram participant A as Server A memory participant N as NIC participant B as Server B memory A->>N: Direct read N->>B: Direct write over the fabric B-->>A: Completion
mindmap
root(("Spine switches"))
n1["Leaf 1"]
n2["Servers"]
n3["Leaf 2"]
n4["Servers"]
n5["Leaf 3"]
n6["Servers"]Key point: Spine-leaf evens out hops.
Why the other options are wrong
- A. A single daisy chain of switches
- A chain adds hops.
- B. A single hub connected by one uplink
- One uplink is a bottleneck.
- C. A ring with one path between neighbors
- One path is a single point of failure.