What is the difference between top-of-rack and end-of-row switching?
Show the answer
Say this first
Top-of-rack puts one switch, or a pair, in every rack. Servers connect with short copper or DAC cables and the switch has fiber uplinks to the spines or aggregation. End-of-row puts a larger switch at the end or middle of a row and runs structured cabling from every rack to it. Spine-leaf normally uses top-of-rack leaves.
Go deeper
| Top-of-rack | End-of-row or middle-of-row | |
|---|---|---|
| Server cables | Short, inside the rack | Long runs across the row |
| Switch count | Many small switches | Few larger chassis |
| Rack as a unit | Yes, pre-cabled racks | No, depends on row cabling |
| Unused ports | Often stranded per rack | Pooled for the row |
| Failure impact | One rack | Whole row |
Practical points
- At 25G and 100G, passive DAC works only for a few meters, which favors top-of-rack. Longer runs need active cables or optics.
- Top-of-rack pairs (MLAG or vPC) let you dual-home every server inside the rack.
- Middle-of-row shortens the longest cable run compared with end-of-row.
They will ask next
When does end-of-row still make sense?
Low-density racks, such as storage or older hardware with a few ports each, where a switch per rack would waste most of its ports.
Traps
- Treating top-of-rack and spine-leaf as the same thing. One is about where the switch sits, the other is the topology.
How did you do?