300-640Updated 8d ago · Aug 28, 2026
Implementing Cisco Data Center AI Infrastructure (DCAI)
60 questions across 1 topics, with suggested answers, explanations where available, and imported community discussion. The first 6 questions are free to preview.
Topics
| # | Topic | Questions | Free |
|---|---|---|---|
| 1 | AI Infrastructure Components and Architecture | 60 | 6 |
Preview
6 of 60 accessibleQuestion 1 · AI Infrastructure Components and Architecture Open
Which components form part of the HyperFabric AI solution?
Answer: C
Cisco Nexus Hyperfabric AI is a cloud-managed, full-stack AI infrastructure that integrates Cisco UCS compute servers with NVIDIA GPUs, Cisco networking, and optional VAST Data storage. Its cloud-hosted controller manages the networking lifecycle.
**Learn more:** [AI Solutions with Cisco Nexus Hyperfabric Data Sheet](https://www.cisco.com/c/en/us/products/collateral/data-center-networking/nexus-hyperfabric/nexus-hyperfabric-ai-ds.html)
Question 2 · AI Infrastructure Operations and Troubleshooting Open
An engineer configures quality of service in a Cisco ACI fabric to connect VAST storage servers. Which attribute combination must be selected?
Answer: A
VAST storage traffic in Cisco ACI uses an RDMA QoS class configured with Weighted Random Early Detection (WRED) and Explicit Congestion Notification enabled. ECN marks packets that would otherwise be dropped, supporting the required RoCEv2 congestion-handling behavior.
**Learn more:** [VAST Data Storage Connectivity to Cisco Nexus 9000 Series Switches in Cisco ACI Mode](https://www.cisco.com/c/en/us/td/docs/dcn/whitepapers/cisco-aci-vast-data.html) · [Cisco APIC and QoS](https://www.cisco.com/c/en/us/td/docs/switches/datacenter/aci/apic/sw/kb/Cisco-APIC-and-QoS.html)
Question 3 · AI Fundamentals and Applications Open
Which IP-based storage transport provides flow control between devices that are directly connected?
Answer: C
NVMe/RoCE runs over RDMA over Converged Ethernet, which uses Priority Flow Control (PFC) for lossless Ethernet service between endpoints. NVIDIA documents that RoCE requires a form of flow control and that PFC is the normal, optimal mechanism.
**Learn more:** [NVIDIA: RDMA over Converged Ethernet (RoCE)](https://docs.nvidia.com/networking/display/mlnxofedv541030/rdma%2Bover%2Bconverged%2Bethernet%2B%28roce%29)