Refer to the exhibit.

The root cause here is likely related to multicast forwarding behavior:

Multicast forwarding may follow a single path based on unicast routing decisions (shortest-path trees or RPF lookup).

Equal-cost Layer 3 links do not always achieve perfect load balancing for multicast traffic since the multicast flow may hash consistently to a particular path.

Applying a more granular load-balancing hash algorithm on SW1 allows better distribution of flows across multiple links by considering additional packet header fields (e.g., Layer 4 ports, source/destination IP).

This is directly aligned with CCDE v3.1 design goals:

Avoiding overutilization by improving hashing entropy.

Preserving consistent unicast and multicast performance.

Leveraging hardware-based load balancing optimizations.

Why other options are incorrect:

A & B: Adding or bundling links doesn't address poor hashing distribution.

D: The issue is identified at SW1 (where hashing decisions are made).

E: Filtering IGMP joins may block legitimate receivers, not a valid solution.

—