mediumMultiple Choice
350-401 Practice Question: An architect is designing a virtualized service…
An architect is designing a virtualized service chain for a campus network using NFV. The chain must include a firewall, WAN optimizer, and IPS. The architect needs to minimize latency by placing VNFs on the same hypervisor host. Which design consideration is most important?
⚠ Common exam trap
Cisco often tests the misconception that DPDK or a Type 2 hypervisor is the primary solution for low-latency NFV, when in fact NUMA awareness is the foundational requirement that must be addressed first.
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
✓
Ensure all VNFs are pinned to the same NUMA node on the hypervisor host.
Pinning all VNFs to the same NUMA node on the hypervisor host minimizes inter-NUMA memory access latency, which is critical for achieving low-latency packet processing in an NFV service chain. When VNFs are placed on different NUMA nodes, memory accesses must traverse the QPI/UPI interconnect, adding significant latency. By co-locating the firewall, WAN optimizer, and IPS on the same NUMA node, the architect ensures that all packet processing stays within the same memory domain, reducing latency to the minimum possible on that host.
Answer analysis
Option-by-option breakdown
For each option: why learners choose it and why it is or isn't the right answer here.
- ✓
Ensure all VNFs are pinned to the same NUMA node on the hypervisor host.
Why this is correct
In multi-socket servers, memory accessed from a remote NUMA node incurs significantly higher latency and consumes inter-socket bandwidth, which can severely degrade throughput for VNF service chains. Pinning all VNFs to the same NUMA node ensures memory allocations are local to the CPU cores executing packet processing, maximizing cache locality and minimizing cross-node traffic. This configuration directly reduces the latency and overhead associated with data plane forwarding between VNFs within a chain.
- ✗
Use a Type 2 hypervisor to reduce overhead.
Why it's wrong here
Type 2 hypervisors sit on top of a host operating system, adding an extra scheduler and abstraction layer that introduces higher CPU overhead and slower I/O processing for virtual machines. In bare-metal NFV environments, Type 1 hypervisors are preferred because they run directly on hardware and provide more deterministic, low-latency performance. The additional overhead from Type 2 architecture would degrade packet throughput and be unsuitable for production virtualized network functions.
- ✗
Place each VNF on a separate physical host to avoid resource contention.
Why it's wrong here
Separating VNFs onto distinct physical hosts eliminates local CPU/memory contention but forces all inter-VNF traffic to traverse physical switches and cables, adding significant network latency and potential congestion. In a service chain, packets must hop multiple times between hosts, which can outweigh any performance gain from avoiding contended resources. Furthermore, this approach reduces resource density and increases network complexity and cost, whereas proper NUMA pinning allows efficient chaining on a single host.
- ✗
Enable DPDK on the virtual switch to accelerate packet processing.
Why it's wrong here
DPDK accelerates packet processing by bypassing the kernel and using busy-polling to achieve high packet rates, but it does not address the physical topology of memory access. If VNFs on different NUMA nodes exchange packets, accessing remote memory still introduces severe latency regardless of DPDK's user-space optimizations. DPDK is a valuable complement to NUMA pinning, not a substitute for it, because the core bottleneck for a contiguous VNF chain is memory locality, not just packet reception speed.
Go deeper
Related to this question
About these practice questions
This 350-401 question is part of Courseiva's 1,923-question bank — original exam-style content with full explanations and wrong-answer analysis, never real exam questions or exam dumps. Learn why practice questions differ from exam dumps →
JA
Written by Johnson Ajibi, MSc IT Security
Senior Network & Security Engineer · founder of Courseiva
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.