Huawei data center design guidelines emphasize end-to-end reliability across all layers of the physical network, including uplinks, devices, interconnections, and server access. All the listed options reflect valid reliability principles.
Option A is correct because ECMP (Equal-Cost Multi-Path) is a fundamental mechanism in spine-leaf architectures, ensuring load balancing and fast convergence when links fail. Option B highlights modular switch reliability , where inter-card connections prevent single-card failures from impacting connectivity.
Option C describes the DAD (Dual-Active Detection) mechanism in M-LAG scenarios. When a peer-link failure occurs, DAD ensures that only one device remains active, preventing split-brain conditions and maintaining network stability.
Option D addresses server-side reliability , where dual-homing with NIC teaming (active-active or active-standby) ensures uninterrupted service in case of link or device failure.
Together, these principles form Huawei’s comprehensive approach to building highly reliable and resilient data center networks.
