300-420

Free 300-420 practice questions

10 questions from our 300-420 practice tests, each with the correct answer and a written explanation. Read them with the answers shown, or switch to quiz mode and answer first.

Back to the 300-420 practice exams

Question 1 of 10, Advanced Addressing and Routing Solutions

A bank is introducing IPv6 in a large network that also has many legacy IPv4-only applications that will not be updated for several years. Users need access to IPv6 internet services immediately, and operations wants one long-term operational model. Which migration strategy is MOST appropriate?

  1. AManual 6in4 tunnels from each branch to a central IPv6 gateway while the core stays IPv4-only
  2. BNative dual stack across the campus and WAN, with IPv4 retired gradually as applications are updated Correct answer
  3. CIPv6-only access with NAT64 and DNS64 at the edge, and the legacy IPv4 servers moved behind NAT64
  4. DISATAP on every host so IPv6 is tunneled across the IPv4 infrastructure without router changes

Answer and explanation: Dual stack gives every host both protocols, so legacy IPv4 applications continue working and IPv6 services are reached natively, and the network converges on a single model as IPv4 is retired. NAT64/DNS64 assists IPv6-only clients reaching IPv4 servers but does not suit IPv4-only clients using legacy applications. Manual tunnels from every branch add overhead and are a transition workaround rather than a long-term model. ISATAP is a host-based tunneling technique that is deprecated in most vendors' design guidance and complicates troubleshooting.

Question 2 of 10, Advanced Addressing and Routing Solutions

An EIGRP design uses summarization at distribution switches toward the core. After a change, the core still sends queries deep into an access block when an access subnet fails. What is the MOST likely design reason?

  1. AThe distribution switches use the wide-metrics EIGRP named mode instead of classic mode
  2. BThe summary is configured on the access-facing interfaces instead of the core-facing interfaces Correct answer
  3. CThe access switches run EIGRP with the default variance of 1, so only successor routes are installed
  4. DThe core routers are configured as EIGRP stub routers with the connected keyword

Answer and explanation: Queries stop at a router that does not have the specific prefix in its topology table, so the summary must be advertised toward the core on the core-facing interfaces; configuring it the other way leaves the core with specific routes and queries continue. Stub configuration on the core would limit queries toward the core rather than away from it and would break transit. Named mode and wide metrics do not affect query boundaries. Variance controls unequal-cost load sharing and is unrelated to query propagation.

Question 3 of 10, Advanced Addressing and Routing Solutions

A service provider-style enterprise backbone runs IS-IS with 220 routers. Engineers want to divide the network so that remote regions keep only intra-region topology and default toward the core, while the core keeps a complete view. What is the BEST IS-IS design?

  1. AMake all routers Level 1-2 and rely on route leaking to keep regional databases small
  2. BMake all routers Level 2 only and configure summarization of each region on the core routers toward one another
  3. CRegional routers as Level 1, regional borders as Level 1-2, and a contiguous Level 2 core Correct answer
  4. DMake regional routers Level 2, the core Level 1, and use the ATT bit for default routing

Answer and explanation: In IS-IS, Level 1 routers keep only their area's topology and use the ATT bit set by L1/L2 routers to install a default route, while the contiguous Level 2 backbone holds inter-area reachability. Making every router L1/L2 maintains two databases everywhere and defeats the goal of small regional databases. An all-L2 design gives every router the full topology. Reversing the roles breaks the model because Level 2 must form the contiguous backbone and the ATT bit is set on L1/L2 routers for Level 1 routers.

Question 4 of 10, Advanced Enterprise Campus Networks

A company with eight SD-Access sites needs to keep macro and micro segmentation end to end between sites over its own WAN, which it controls. Which SD-Access transit type should be used?

  1. ALayer 2 border handoff, which extends fabric VLANs to traditional switches at each site
  2. BInternet transit using Cisco Secure Client tunnels from each fabric edge to the hub
  3. CIP-based transit, which hands traffic off as VRF-Lite and requires SGT reapplication
  4. DSD-Access transit, which keeps VXLAN encapsulation and SGT information between sites Correct answer

Answer and explanation: SD-Access transit keeps the VXLAN encapsulation, VN and SGT across sites, using transit control plane nodes, so end-to-end segmentation is preserved over a controlled WAN. IP-based transit decapsulates at the border and requires VRF-Lite plus separate SGT propagation such as inline tagging or SXP. A Layer 2 border handoff connects a fabric to a traditional VLAN, not to other sites. Client VPN tunnels are a remote access solution and not a fabric transit.

Question 5 of 10, Advanced Enterprise Campus Networks

A company deploys fabric-enabled wireless in SD-Access. Where is the client data traffic from the access points switched?

  1. AThe AP encapsulates client traffic in GRE to the control plane node for policy lookup
  2. BThe AP sends client traffic over FlexConnect local switching to the access VLAN
  3. CThe AP sends client traffic in VXLAN to its fabric edge node, which forwards it in the fabric Correct answer
  4. DThe AP sends client traffic inside CAPWAP to the WLC, which then forwards it toward the fabric border node

Answer and explanation: With fabric-enabled wireless, the WLC handles control functions via CAPWAP, while APs send client data in VXLAN to the connected fabric edge, so wireless traffic gets the same distributed forwarding and policy as wired traffic. Sending data through CAPWAP to the WLC is the over-the-top model. FlexConnect local switching is a different deployment model. The control plane node does not forward data traffic.

Question 6 of 10, WAN for Enterprise Networks

A company is sizing Cisco SD-WAN controllers for 1,800 branches with two transports each. Which design consideration is correct for controller redundancy?

  1. ADeploy a single SD-WAN Manager cluster member, which also acts as Controller and Validator
  2. BMultiple Controllers and Validators in different locations, with controller groups or affinity Correct answer
  3. CDeploy one Controller per data center and configure each edge to connect to only one of them
  4. DDeploy Controllers only in the primary data center, because edges keep forwarding with cached state during outages

Answer and explanation: Redundant Controllers and Validators in geographically separate locations, with controller groups or affinity, distribute the load and survive a site failure. Connecting each edge to only one Controller removes redundancy. SD-WAN Manager does not replace the Controller and Validator roles, and a single node is not redundant. Edges continue forwarding for a period using existing state, but losing all Controllers eventually affects keys and route updates, so controllers should not be in one location only.

Question 7 of 10, WAN for Enterprise Networks. Choose 2.

A company is designing Cisco SD-WAN for 2,000 sites. Which TWO controller design practices are correct? (Choose two.)

  1. AUse one Validator because it does not need redundancy
  2. BDeploy Controllers in multiple data centers or cloud regions Correct answer
  3. CPlace the Manager in VPN 512 of every edge router
  4. DSize the number of Controllers to the expected edge and tunnel count Correct answer
  5. EConnect each edge to every Controller in the overlay

Answer and explanation: Geographic distribution of Controllers protects against site failures, and the number of Controllers must be sized to the number of edges and control connections. Validators also need redundancy, often through DNS with multiple addresses. Manager is not placed in edge router VPNs. Each edge connects to a limited number of Controllers, governed by its control-connection limit and controller affinity settings, not to all of them.

Question 8 of 10, Network Services

A WAN interface uses low-latency queuing. During a large video conference, critical business traffic starves even though its class has a bandwidth guarantee. Investigation shows the video is marked EF and placed in the priority queue. What design change is MOST appropriate?

  1. AMove interactive video to its own AF41 class with a bandwidth guarantee, keep EF for voice, and cap priority at about one third of the link Correct answer
  2. BIncrease the priority queue to 80% of the link so that all voice and video traffic marked EF is protected during the conference
  3. CEnable WRED on the business class so that it drops traffic earlier during congestion
  4. DRemove the policer from the priority queue so that EF traffic is never dropped

Answer and explanation: Cisco guidance limits strict priority traffic to roughly 33% of the link and recommends separating video into its own class or a policed priority level, otherwise large real-time loads starve other classes. Increasing the priority queue to 80% worsens starvation. The priority queue is implicitly policed during congestion to protect other traffic, so removing that protection is harmful. WRED on the business class would make it drop earlier and does not fix the starvation.

Question 9 of 10, Network Services

A branch has a 1-Gbps Ethernet handoff to a provider, but the contracted service is 200 Mbps. The provider drops excess traffic. What must the branch router do so that its queuing policy can protect voice and critical applications?

  1. AMark all traffic as best effort so the provider applies its own fair queuing
  2. BApply a hierarchical policy that shapes to 200 Mbps with a child queuing policy Correct answer
  3. CApply a queuing policy directly to the physical 1-Gbps interface without shaping
  4. DApply a policer at 200 Mbps on the outbound interface and drop excess traffic

Answer and explanation: Because the physical link never congests, queuing does not engage unless the router shapes to the contracted rate; a parent shaper with child queuing creates congestion at the router where priority is controlled. A policer drops packets without prioritization. Queuing on the interface without shaping never activates. Remarking to best effort removes differentiation and leaves the provider dropping traffic arbitrarily.

Question 10 of 10, Automation

An automation team must configure a Cisco-specific feature on Catalyst 9000 switches that is not represented in IETF or OpenConfig models. The team still wants model-driven programmability. Which approach is correct?

  1. AUse the IETF interface models and encode the feature in the description leaf
  2. BUse SNMP SET operations against the MIB because YANG models are not usable
  3. CUse the OpenConfig models and add the missing settings through gNMI extensions
  4. DUse the Cisco IOS XE native YANG models, which cover most platform features Correct answer

Answer and explanation: Cisco native models map closely to the IOS XE configuration and expose platform-specific features that vendor-neutral models do not cover. Extensions in gNMI do not add missing schema to OpenConfig. Encoding configuration in a description leaf does not configure anything. SNMP SET is not model-driven programmability in the YANG sense and lacks many configuration objects.