You have a customer using Spine and Leaf design with two 8325 32-port 100G QSFP28 (F-B airflow) Spine Switches. What is the maximum number of racks supported if each Rack has two 8360 VSX stacked…

Question

Options

• AUp to 16 total racks using different subscription rates
• BUp to 4 total racks using the same subscription rates
• CUp to 4 total racks using different subscription rates
• DUp to 8 total racks using the same subscription rates

• A
  23% (7)
• B
  6% (2)
• C
  10% (3)
• D
  61% (19)

Explanation

Option D is correct because each 8325 spine has 32 x 100G ports, and each rack requires 4 spine ports (2 DACs × 2 leaf switches per VSX pair). Simple division gives 32 ÷ 4 = 8 racks per spine. Since both spines are identical and every rack connects symmetrically to both spines, all racks receive equal uplink bandwidth - meaning the same subscription rate across the fabric.

Why the distractors fail:

• A (16 racks, different rates): Reaching 16 would require only 2 ports per rack per spine, ignoring that each rack has two leaf switches each needing 2 DACs - that's 4 ports per rack, not 2.
• B (4 racks, same rates): The math for 4 racks would mean only 1 DAC per leaf switch, contradicting the stated 2x DAC design.
• C (4 racks, different rates): Wrong on both counts - 4 is half the correct number, and the symmetric dual-spine design inherently produces equal subscription rates.

Memory tip: Think "2×2=4, 32÷4=8." Two spines, two leaf switches per rack, two DACs per leaf - collapse the per-rack port cost to 4, then divide the 32-port spine by 4. That "same" subscription rate is the giveaway: asymmetry only arises when you oversubscribe some racks differently than others, which a clean symmetric design avoids.

No community discussion yet for this question.