Mention two disadvantages of the End-of-Row or Middle-of-Row switches in Data Center design

Question

Options

• AMore equipment
• BHigher latency for rack-to-rack traffic within the row
• CLonger, more complex cabling within the row
• DLower port density and less scalable

• A
  7% (3)
• B
  4% (2)
• C
  89% (41)

Explanation

C and D are correct because EoR/MoR designs place a shared aggregation switch at the row's end or middle, forcing cables from every rack in the row to traverse the full horizontal length to reach it - creating longer, more complex horizontal cable runs (C). That same shared switch must provision ports for all racks simultaneously, which caps how many devices each rack can connect and makes expanding the row constrained by a single switch's fixed port count (D).

A is wrong - EoR/MoR actually uses fewer switches than Top-of-Rack (ToR), where each rack has its own dedicated switch. Reduced equipment count is an advantage of EoR/MoR.

B is wrong - intra-row rack-to-rack traffic in EoR/MoR traverses only one switch hop (the shared row switch). In ToR designs, the same traffic must traverse two ToR switches plus an aggregation switch - typically more latency, not less.

Memory tip: Think "End-of-Row = End of your cable budget." All cables must travel to one end, making them longer and messier (C), and one switch for the whole row means fewer ports per rack and painful scaling (D). The things EoR saves (equipment count, latency) are the wrong-answer distractors.

No community discussion yet for this question.