H19-401_V2.0 practice question 4 of 15
Distributed VXLAN gateways are recommended if VXLAN networking is required.
Full set: 60 questionsThe statement is correct. In modern data center networks utilizing VXLAN, a distributed gateway architecture is the highly recommended and standard deployment model. In this design, the Layer 3 gateway functionality is distributed to all leaf switches (VTEPs) in the fabric. This allows for optimal traffic forwarding, particularly for East-West (server-to-server) traffic between different subnets. Traffic is routed at the first-hop leaf switch, eliminating the inefficient "traffic tromboning" effect where traffic must first travel to a centralized gateway before reaching its destination within the same fabric. This approach significantly reduces latency, improves bandwidth utilization, enhances scalability, and increases overall network resilience.
The alternative to a distributed gateway is a centralized gateway model. While functionally viable, it is not the recommended approach for scalable VXLAN deployments. A centralized model creates a potential performance bottleneck, as all inter-subnet and North-South traffic must traverse a single point. This increases latency for East-West traffic and limits the overall scalability of the network. Therefore, stating that distributed gateways are not recommended (the "False" option) would be incorrect as it contradicts modern data center networking best practices.
1. Huawei Official Vendor Documentation: In the Huawei CloudFabric 3.0 Hyper-Converged Data Center Network Solution Technical White Paper, the architecture is explicitly based on a spine-leaf topology with distributed VXLAN gateways. The paper states, "The solution uses the industry's mainstream spine-leaf architecture... On this network, distributed VXLAN gateways are deployed, and the forwarding plane uses VXLAN." This document emphasizes that this distributed model enables "one-hop forwarding of traffic on the entire network," which is a key advantage. (Reference: Huawei CloudFabric 3.0 Hyper-Converged Data Center Network Solution Technical White Paper, Section 2.1 "Spine-Leaf Architecture").
2. Huawei Official Vendor Documentation: The Huawei CloudFabric Easy Solution Technical White Paper also describes the benefits of the distributed model. It explains that with distributed gateways on leaf nodes, "inter-subnet traffic between VMs on a rack is forwarded through the leaf node," which "implements optimal traffic forwarding and prevents traffic detours." (Reference: Huawei CloudFabric Easy Solution Technical White Paper, Section "Distributed VXLAN Gateway").
3. University Courseware: Stanford University's course CS244, "Advanced Topics in Networking," discusses modern data center architectures. Lecture materials often cover the transition from traditional hierarchical networks to spine-leaf fabrics. The principles taught explain how distributing routing and gateway functions to the edge (leaf switches) using protocols like VXLAN solves the scalability and East-West traffic challenges inherent in older, centralized designs. The distributed model is presented as the solution for efficient, scalable data center operations. (Reference: Concepts covered in Stanford CS244 curriculum on Data Center Networking).
Keep practicing
Take the free H19-401_V2.0 practice test
Ten exam-style questions with answers and explanations, plus the exam facts and study guides.
More questions
Other H19-401_V2.0 practice questions
- Question 1Which of the following views is not one of the three views of iMaster NCE-CampusInsight?
- Question 2iMaster NCE-CampusInsight allows for capacity optimization by identifying weak-coverage A…
- Question 3On a traditional network, network planning, configuration, and O&M are performed by profe…
- Question 5iMaster NCE-Campus has a built-in terminal fingerprint database for identifying terminals.
- Question 6When iMaster NCE-Campus functions as an authentication server, it can connect to common u…
- Question 7In the SD-WAN Solution, the RR and CPE cannot be the same device.
