You need to recommend a solution that meets the following requirements: - Minimizes the use of the virtual machine processors to transfer data - Minimizes network latency Which virtual machine size…

Question

Exhibits

Options

• ACompute optimized Standard_F8s
• BGeneral purpose Standard_B8ms
• CHigh performance compute Standard_H16r
• DMemory optimized Standard_E16s_v3
• EReceive side scaling (RSS)
• FRemote Direct Memory Access (RDMA)
• GSingle root I/O virtualization (SR-IOV)
• HVirtual Machine Multi-Queue (VMMQ)

• A
  2% (1)
• B
  3% (2)
• C
  73% (44)
• D
  13% (8)
• E
  7% (4)
• G
  2% (1)

Explanation

Standard_H16r is Azure's High Performance Compute VM with the r suffix explicitly indicating built-in RDMA support. RDMA (Remote Direct Memory Access) transfers data directly between machines' memory, completely bypassing the CPU and OS networking stack - which simultaneously minimizes processor usage and network latency, satisfying both requirements.

Why the distractors are wrong:

• A (F8s) and D (E16s_v3): Compute/memory-optimized VMs lack RDMA support; their network is standard.
• B (B8ms): Burstable general-purpose VMs are designed for variable workloads, not low-latency networking.
• E (RSS): Receive Side Scaling distributes network interrupts across multiple CPU cores - it actually increases CPU involvement, not decreases it.
• G (SR-IOV): Enables Accelerated Networking (reduces latency via hardware passthrough), but still routes through the CPU; it doesn't achieve the zero-CPU-copy that RDMA does.
• H (VMMQ): Like RSS, it uses multi-queue processing across CPUs - more CPU involvement, not less.

Memory tip: Think "RDMA = RAM-to-RAM, no CPU in between." The r in H16r stands for RDMA - whenever an exam question says minimize CPU for data transfer, the answer is always RDMA + an HPC VM that supports it.

No community discussion yet for this question.