mediumMultiple Select
350-401 Practice Question: Which two statements about NFV performance…
Which two statements about NFV performance considerations are true? (Choose two.)
⚠ Common exam trap
The trap here is that NUMA awareness sounds like a fundamental performance technique, so candidates pick it over the more specific SR-IOV/DPDK pair; the exam expects you to distinguish primary data-plane accelerators from supporting tuning knobs.
Answer choices
Why each option matters
Answer the question above first, then reveal the full breakdown to understand why each option is right or wrong.
Correct answer & explanation
✓
SR-IOV allows a virtual function (VF) to be directly assigned to a VM, providing near-native network performance.
Option A is correct because SR-IOV (Single Root I/O Virtualization) lets a physical NIC expose virtual functions (VFs) that can be directly assigned to a VM via PCI passthrough, so traffic bypasses the hypervisor's virtual switch and gives near-native throughput and low latency. Option C is correct because DPDK (Data Plane Development Kit) supplies user-space libraries, PMD drivers, and hugepage-based memory management that let applications poll NICs directly, bypassing the kernel network stack to achieve high packet-processing rates. Option B is not correct as stated because NUMA awareness is an important optimization for memory locality, but it is not the primary technique for NFV packet processing; SR-IOV, DPDK, and CPU pinning play more central roles. Option D is wrong because bonding multiple uplinks improves redundancy and aggregate bandwidth but does not remove the need for tuning such as DPDK, SR-IOV, or CPU pinning. Option E is wrong because NFV performance can be substantially improved with the right hardware and software optimizations, often approaching or matching physical appliance performance.
Answer analysis
Option-by-option breakdown
For each option: why learners choose it and why it is or isn't the right answer here.
- ✓
SR-IOV allows a virtual function (VF) to be directly assigned to a VM, providing near-native network performance.
Why this is correct
SR-IOV bypasses the hypervisor's virtual switch by assigning a hardware virtual function directly to the guest, eliminating software-based packet processing overhead. This satisfies the stem's performance constraint, delivering near-native throughput and reduced CPU load compared with virtio or full emulation, which matters for NFV workloads requiring line-rate forwarding.
- ✗
NUMA awareness is the primary technique to improve NFV packet processing performance.
Why it's wrong here
NUMA awareness reduces cross-socket memory latency but is one of several techniques; DPDK, SR-IOV, CPU pinning and hugepages contribute at least as much to packet processing throughput. It is tempting because NUMA placement genuinely matters on multi-socket servers, yet it is not the primary lever for NFV performance.
- ✓
DPDK provides a set of libraries and drivers for fast packet processing in user space, bypassing the kernel network stack.
Why this is correct
DPDK satisfies the NFV requirement for high-throughput packet processing by moving data-plane workloads into user space. Poll-mode drivers and hugepage-backed memory eliminate kernel interrupts and copies, cutting per-packet overhead. This directly addresses the performance bottleneck virtualised network functions face when relying on the standard Linux kernel stack.
- ✗
Using a virtual switch with multiple bonded uplinks eliminates the need for any performance optimization.
Why it's wrong here
Bonded uplinks provide link redundancy and aggregate bandwidth, but they do not remove CPU, memory-bandwidth or vSwitch datapath bottlenecks that limit NFV throughput. It is tempting because bonding appears to add capacity, and it would be the right answer for a question about link resilience rather than performance optimisation.
- ✗
NFV performance is inherently lower than physical appliances and cannot be improved.
Why it's wrong here
NFV throughput is constrained by kernel datapath, vSwitch and CPU pinning choices, not an immutable property; DPDK, SR-IOV and hardware offload demonstrably raise packet rates toward line rate. It is tempting because early NFV deployments did underperform, so the claim sounds like accepted wisdom rather than an absolute limitation.
Go deeper
Related to this question
Learn chapter
QoS and Network Performance Management
Key term
Virtual Machine Networking
Virtual machine networking is how virtual computers on a physical server connect to each other, to the internet, and to the rest of a network.
Key term
Virtual Switch
A virtual switch is a software-based network switch that connects virtual machines within a hypervisor and forwards traffic between them and the physical network.
About these practice questions
One of 1,923 original 350-401 practice questions on Courseiva, each with a full explanation and wrong-answer analysis — not exam dumps or protected exam content. Learn why practice questions differ from exam dumps →
JA
Written and reviewed by Johnson Ajibi, MSc IT Security
Senior Network & Security Engineer · founder of Courseiva
Last reviewed September 2026 · checked against the official Cisco exam blueprint
This 350-401 practice question is part of Courseiva's free Cisco certification practice question bank. Courseiva provides original exam-style practice questions with explanations, topic-based practice, mock exams, readiness tracking, and study analytics to help learners prepare for the 350-401 exam.