The correct answer is A. Change the hash algorithm. In a LAG, traffic is not normally split packet-by-packet across every member link. Instead, the switch uses a hashing method to choose which physical interface in the aggregate should transmit each flow. The output shows that the current hash is l3-src-dst, which means the load-balancing calculation uses Layer 3 source and destination IP information. If many flows share the same source and destination IP pair, traffic can be unevenly distributed, causing one link to become congested while another link is lightly used. HPE Aruba Networking AOS-CX documentation states that the lacp hash command controls the selection of an interface in an aggregate group and supports Layer 2, Layer 3, and Layer 4 source/destination hash options. Changing the hash algorithm, for example to include Layer 4 source and destination TCP/UDP ports, can improve distribution when Layer 3 values alone are not diverse enough. Changing LACP mode, configuring static LAG, or changing LACP heartbeat rate would not directly fix uneven traffic distribution.