Switch fabric design: In a cell-backplane architecture, should the internal backplane throughput be significantly higher than the sum of the port data rates to avoid bottlenecks?

Introduction:
Modern switches often use a fabric or cell-based backplane to move traffic between ingress and egress ports. A key design metric is whether the internal switching capacity (backplane bandwidth) can handle worst-case traffic patterns without becoming a chokepoint. This question tests understanding of non-blocking switch design principles.

Given Data / Assumptions:

• Multiple ports can send/receive simultaneously.
• Traffic patterns can be adversarial (many-to-one, incast) or uniform.
• The backplane aggregates flows and must avoid internal congestion.

Concept / Approach:
A “non-blocking” or “wire-speed” switch must ensure its internal fabric capacity meets or exceeds aggregate port demand, factoring full-duplex operation (ingress + egress) and oversubscription targets. In cell-backplane designs, frames are segmented into cells and switched across a crossbar or fabric. Provisioning a backplane rate significantly higher than the sum of ports reduces head-of-line blocking and ensures sustained throughput under load.

Step-by-Step Solution:
1) Sum potential port throughput (consider full-duplex).2) Compare to internal fabric capacity.3) If backplane < aggregate, internal contention limits performance.4) Therefore, design for backplane ≥ aggregate (often greater) to avoid bottlenecks.

Verification / Alternative check:
Vendor datasheets quote “switching capacity” and “forwarding rate.” Non-blocking models list capacities at or above aggregate line rate for all ports concurrently.

Why Other Options Are Wrong:

• Incorrect / only for half-duplex / only with jumbo / only for fiber: Backplane provisioning concerns exist regardless of duplex, frame size, or medium.

Common Pitfalls:
Comparing only uplink bandwidth and ignoring internal crossbar limits; missing that full-duplex doubles edge demand.

Final Answer:
Correct