Which statement about iSCSI RDMA is true?

• A. iSCSI RDMA uses RDMA over Ethernet to reduce CPU overhead.
• B. iSCSI RDMA requires Fibre Channel fabric.
• C. iSCSI RDMA cannot be used with TCP/IP.
• D. iSCSI RDMA is slower than standard iSCSI.

Answer: (A)

iSCSI RDMA moves data using RDMA transports, which lets data flow directly between memory regions with minimal CPU involvement. When deployed over Ethernet with RoCE or over an InfiniBand fabric, this offloads a lot of the work from the host CPU, reducing CPU overhead, lowering latency, and boosting throughput compared to standard iSCSI that runs data through the regular TCP/IP stack. That is why the statement about using RDMA over Ethernet to cut CPU overhead is the correct one. It doesn’t require Fibre Channel—FC is a separate storage fabric, not a prerequisite for iSCSI RDMA. The notion that it cannot be used with TCP/IP isn’t accurate in practice, because the data path uses RDMA while control and management can still rely on IP/TCP in many deployments. In short, RDMA-enabled iSCSI delivers the performance benefits by bypassing the CPU-heavy data path of traditional iSCSI.