I85PREP
⌘ K
JNCDS-DC study materials
JNCDS-DC study resources

JNCDS-DC practice questions

One question at a time. Build your JNCDS-DC understanding.

First 50 of 200 questions. Each one lists the choices, the correct answer, and an explanation. The checker at the top of this page saves progress in this browser when it is available.

Continue in the practice tool above to load more questions, 50 at a time.

Question 1Data Center Architecture and Design Principles

In a Junos-based data center, which protocol is the standard control-plane mechanism for dynamically signaling VXLAN tunnel endpoints between leaf and spine devices?

  1. STP with RSTP extensions
  2. LACP over aggregated Ethernet bundles
  3. OSPFv2 with point-to-point links
  4. BGP with the EVPN address family Correct answer.

Explanation

EVPN uses BGP to distribute MAC and IP reachability information, so leaf and spine devices learn remote VXLAN tunnel endpoints and tenant MAC addresses without data-plane flooding. OSPF, LACP, and STP operate at other layers and cannot advertise MAC or VTEP reachability.

Question 2Data Center Architecture and Design Principles

A design team is collapsing the traditional three-tier data center into a spine-leaf fabric. Which characteristic is the primary architectural reason for this change?

  1. It eliminates the need for routing protocols in the underlay
  2. It guarantees deterministic latency of under one microsecond between any two hosts
  3. It provides predictable any-to-any paths with consistent hop counts between leaves Correct answer.
  4. It removes the requirement for redundant power in the data center

Explanation

A spine-leaf topology gives every leaf an equal-cost path to every other leaf through the spines, so east-west traffic sees a consistent and predictable hop count. This predictability is the main driver for replacing oversubscribed three-tier designs. Sub-microsecond latency, protocol-free underlay, and power redundancy are not properties of the topology itself.

Question 3Data Center Architecture and Design Principles

Which statement correctly describes the role of an IP fabric underlay in an EVPN-VXLAN data center design?

  1. It replaces EVPN by learning MAC addresses in hardware
  2. It provides routed reachability between loopback addresses of the VTEP devices Correct answer.
  3. It aggregates all tenant traffic onto a single pair of core switches
  4. It carries tenant Layer 2 frames natively between all servers

Explanation

The underlay is a routed IP fabric whose job is to reach the loopback address of each VTEP, typically using ECMP across the spines. Tenant Layer 2 and Layer 3 services ride in the VXLAN overlay on top of that routed transport.

Question 4Data Center Architecture and Design Principles

What is the purpose of a VTEP in a VXLAN data center fabric?

  1. It terminates storage-area-network Fibre Channel traffic
  2. It maps tenant VLANs to VNIs and encapsulates frames in VXLAN tunnels Correct answer.
  3. It assigns DHCP addresses to bare-metal servers in the fabric
  4. It provides out-of-band management access to spine switches

Explanation

A VXLAN Tunnel Endpoint sits at the edge of the overlay, mapping local VLANs to VNIs and performing encapsulation and decapsulation of tenant frames. On Junos devices this role is typically performed by leaf switches configured with VXLAN and EVPN.

Question 5Data Center Architecture and Design Principles

In a collapsed spine-leaf design, a server is dual-homed to two leaf switches using a multichassis LAG. Which Junos feature provides active-active multichassis link aggregation between those two leaves?

  1. Multichassis LAG (MC-LAG) with ICCP Correct answer.
  2. Virtual Chassis Fabric with a single control plane
  3. Redundant trunk group with VRRP
  4. Aggregated Ethernet with LACP only

Explanation

MC-LAG on Junos uses Inter-Chassis Control Protocol to coordinate the two leaf devices so they present an active-active LAG to the dual-homed server. ICCP synchronizes state so both peers agree on which links may forward.

Question 6Data Center Architecture and Design Principles

A design requires that a dual-homed server's default gateway MAC address remain stable if one of its two leaf switches fails. Which approach achieves this in an EVPN-VXLAN fabric?

  1. Configuring a unique VRRP master on each leaf with different virtual MACs
  2. Assigning the server a static route to each leaf loopback
  3. Relying on proxy ARP alone on both leaf switches
  4. Using an anycast gateway with the same IP and MAC configured on both leaves Correct answer.

Explanation

An anycast (virtual) gateway configures the identical gateway IP and MAC on every leaf, so the server's first-hop information never changes when a leaf fails. Traffic is then routed locally by whichever leaf is reachable. Different VRRP virtual MACs or static routes would break that stability.

Question 7Data Center Architecture and Design Principles

When designing the underlay for a large EVPN-VXLAN fabric, which addressing choice best supports stable tunnel endpoint reachability?

  1. Rely on link-local IPv6 addresses for VTEP peering
  2. Use dedicated loopback interfaces advertised into the underlay routing protocol Correct answer.
  3. Advertise VTEP addresses from physical interface subnets that may flap
  4. Use the management interface addresses as tunnel endpoints

Explanation

VTEP addresses should be stable loopbacks reachable through the routed fabric, independent of any single physical link state. Advertising physical interface subnets risks losing reachability when a link goes down, and management addresses are not part of the data-plane fabric.

Question 8Data Center Architecture and Design Principles

What is the primary benefit of using an ESI (Ethernet Segment Identifier) in an EVPN-VXLAN design with dual-homed servers?

  1. It encrypts all traffic between the server and the leaf switches
  2. It assigns the server a public IP address automatically
  3. It allows the fabric to treat the multihomed server links as a single logical segment for forwarding and aliasing Correct answer.
  4. It replaces the need for a routing protocol in the underlay

Explanation

An ESI identifies a set of links that belong to the same multihomed Ethernet segment, enabling EVPN to perform designated forwarder election and aliasing so traffic is balanced and loops are avoided. It is a control-plane identifier, not an encryption or addressing mechanism.

Question 9Data Center Architecture and Design Principles

Which Junos routing instance type is typically used to provide Layer 3 tenant separation in an EVPN-VXLAN fabric?

  1. mpls LSP instances
  2. forwarding-only instances without a route table
  3. virtual-router routing instances with VRF tables Correct answer.
  4. logical-system instances tied to physical interfaces

Explanation

A virtual-router routing instance provides a separate VRF table for each tenant, and EVPN type-5 (or symmetric IRB type-2) routes distribute tenant prefixes across the fabric. Forwarding-only instances lack the routing table needed for tenant routing.

Question 10Data Center Architecture and Design Principles

A designer wants to prevent unknown unicast flooding across the EVPN-VXLAN overlay. Which design element most directly supports this?

  1. Increasing the MTU on all spine links
  2. Advertising MAC and IP routes through EVPN so leaves learn remote addresses Correct answer.
  3. Disabling ARP suppression on all leaf switches
  4. Enabling storm control on every server port

Explanation

When EVPN advertises MAC and IP reachability, leaves learn remote hosts from the control plane and can forward known unicast directly, avoiding flooding. ARP suppression further reduces broadcast and unknown traffic. Storm control and MTU changes do not eliminate unknown unicast flooding.

Question 11Data Center Architecture and Design Principles

Which statement about oversubscription ratio in a spine-leaf fabric is correct?

  1. A lower oversubscription ratio generally increases available east-west bandwidth per server Correct answer.
  2. Higher oversubscription always improves application latency
  3. Oversubscription only applies to the management network
  4. Oversubscription is eliminated entirely by using VXLAN

Explanation

Oversubscription ratio compares host-facing bandwidth to uplink bandwidth; lowering it provides more east-west capacity per server, which matters for distributed applications. VXLAN does not change physical oversubscription, and higher ratios reduce available bandwidth.

Question 12Data Center Architecture and Design Principles

In a Junos EVPN-VXLAN deployment, what does symmetric IRB provide compared with asymmetric IRB?

  1. Routing occurs only on the ingress leaf
  2. It forces all inter-tenant traffic through a central firewall
  3. It eliminates the need for VNIs in the overlay
  4. Both ingress and egress leaves route the packet using their local VRF tables Correct answer.

Explanation

Symmetric IRB routes at both the ingress and egress leaf, with each leaf holding the tenant VRF, which scales better and keeps policy consistent. Asymmetric IRB routes only at ingress and relies on the egress leaf having the destination subnet locally.

Question 13Data Center Architecture and Design Principles

A fabric uses EVPN with a designated forwarder election on a multihomed Ethernet segment. What is the purpose of this election?

  1. To decide which leaf performs DHCP relay
  2. To elect the BGP route reflector for the fabric
  3. To select one PE to forward broadcast, unknown unicast, and multicast traffic toward the segment Correct answer.
  4. To choose which spine carries all inter-leaf unicast traffic

Explanation

Designated forwarder election ensures only one of the multihomed PEs forwards BUM traffic into a given Ethernet segment, preventing duplicate delivery. Unicast forwarding and BGP reflector roles are handled by other mechanisms.

Question 14Data Center Architecture and Design Principles

Which design consideration most affects the choice between a Layer 2 stretched fabric and a pure Layer 3 fabric for a new data center?

  1. Whether workloads require live migration across subnets or legacy Layer 2 adjacency Correct answer.
  2. The color coding standard for patch cables
  3. The number of power feeds available in each rack
  4. The brand of optical transceivers used on spine uplinks

Explanation

Layer 2 extension is typically driven by applications that need subnet-preserving live migration or legacy Layer 2 dependencies. A pure Layer 3 design is simpler and more scalable when those constraints do not exist. Rack power, optics vendors, and cabling conventions do not determine the Layer 2 versus Layer 3 architecture.

Question 15Data Center Architecture and Design Principles

What role does a route reflector serve in an EVPN-VXLAN spine-leaf fabric?

  1. It performs MAC address aging for the overlay
  2. It assigns VNIs to VLANs on each leaf
  3. It forwards all tenant data traffic between leaves
  4. It reflects EVPN routes between leaf peers so a full iBGP mesh is unnecessary Correct answer.

Explanation

Spines commonly act as BGP route reflectors, distributing EVPN routes among leaves without requiring every leaf to peer with every other leaf. Data forwarding remains on the leaf-spine path, and VNI mapping is a local configuration task.

Question 16Data Center Architecture and Design Principles

A design uses EVPN type-5 routes. What information do these routes primarily carry?

  1. Ethernet segment identifiers for multihoming
  2. Multicast group membership for underlay replication
  3. MAC addresses only, with no IP prefix information
  4. IP prefix reachability for tenant subnets across the fabric Correct answer.

Explanation

EVPN type-5 (IP prefix) routes advertise tenant IP subnets so that Layer 3 traffic can be routed across the fabric without requiring every subnet on every leaf. Type-2 routes carry MAC and IP host information, and type-4 routes carry ESI information.

Question 17Data Center Architecture and Design Principles

Which statement best describes the function of ARP suppression in an EVPN-VXLAN leaf?

  1. The leaf converts ARP into DHCP requests
  2. The leaf blocks all ARP traffic on the overlay
  3. The leaf answers ARP requests for known remote hosts from its EVPN-learned table Correct answer.
  4. The leaf floods ARP requests to all spines

Explanation

With ARP suppression, the leaf responds to ARP requests for hosts it has learned through EVPN, reducing broadcast flooding across the overlay. It does not block ARP entirely or transform it into another protocol.

Question 18Data Center Architecture and Design Principles

In a data center fabric design, what is the main purpose of ECMP on the spine layer?

  1. To provide Layer 2 loop prevention
  2. To encrypt traffic between leaves
  3. To elect a single active spine for all traffic
  4. To distribute traffic across multiple equal-cost paths toward the destination leaf Correct answer.

Explanation

ECMP lets a leaf spread traffic across all available spines when multiple equal-cost paths exist, improving utilization and resilience. It does not elect a single active spine, prevent Layer 2 loops, or provide encryption.

Question 19Data Center Architecture and Design Principles

Which factor most directly determines the number of spines needed in a non-blocking spine-leaf fabric?

  1. The number of racks in the facility
  2. The speed of the management network
  3. The number of VLANs configured on servers
  4. The total leaf uplink bandwidth that must be absorbed at the spine layer Correct answer.

Explanation

Spine count is sized so the aggregate leaf uplink bandwidth can be absorbed while maintaining the target oversubscription, which depends on port count and port speed at the spine. Rack count, VLAN count, and management speed are not the direct sizing driver.

Question 20Data Center Architecture and Design Principles

What does a VNI identify in a VXLAN overlay?

  1. A redundant power supply unit
  2. A Layer 2 or Layer 3 segment in the overlay Correct answer.
  3. A BGP autonomous system number
  4. A physical uplink port on a spine

Explanation

A VNI is a 24-bit identifier that maps to an overlay segment, allowing the same VLAN ID to be reused independently per tenant. It is not a physical port, an AS number, or a hardware identifier.

Question 21Data Center Architecture and Design Principles

A design team wants to connect two data centers with EVPN-VXLAN while keeping each site's underlay independent. Which approach is most appropriate?

  1. Use EVPN over a Layer 3 interconnect such as a WAN or DCI link between route reflectors or border leaves Correct answer.
  2. Merge both sites into one Virtual Chassis
  3. Stretch a single VLAN across the interconnect for all tenants
  4. Extend the same underlay IGP across both sites as one area

Explanation

EVPN can span sites over a Layer 3 interconnect, with border devices or route reflectors exchanging EVPN routes while each site keeps its own underlay. Extending one IGP or stretching VLANs across sites increases failure domain and complexity.

Question 22Data Center Architecture and Design Principles

Which Junos configuration element associates a VLAN with a VXLAN VNI on a leaf switch?

  1. A virtual-switch or bridge-domain mapping that binds the VLAN to the VNI Correct answer.
  2. A firewall filter applied to the management interface
  3. A class-of-service scheduler map
  4. An OSPF area assignment on the loopback

Explanation

On Junos, VLANs are bound to VNIs through bridge-domain or virtual-switch style configuration so the VTEP knows which overlay segment carries the frames. OSPF areas, firewall filters, and CoS maps serve unrelated functions.

Question 23Data Center Architecture and Design Principles

What is the main benefit of a Layer 3 fabric up to the server (routed server access) compared with Layer 2 server access?

  1. It forces all traffic through a single default gateway
  2. It requires spanning tree to prevent loops
  3. It eliminates large broadcast domains and improves multipath utilization Correct answer.
  4. It removes the need for IP addressing on servers

Explanation

Routing at the server edge removes large Layer 2 broadcast domains and lets ECMP use all available paths, improving scalability and convergence. Servers still need IP addressing, and spanning tree is not required in a fully routed design.

Question 24Data Center Architecture and Design Principles

Which statement about the relationship between MTU and VXLAN encapsulation is correct?

  1. VXLAN reduces frame size, so the underlay MTU can be lowered
  2. VXLAN encapsulation adds overhead, so the underlay MTU must be increased to avoid fragmentation Correct answer.
  3. MTU has no effect on VXLAN traffic
  4. VXLAN requires the underlay MTU to match the management network MTU exactly

Explanation

VXLAN adds outer headers, so the underlay must support a larger MTU than the tenant payload to prevent fragmentation or drops. Reducing MTU or ignoring the overhead leads to fragmentation and performance problems.

Question 25Data Center Architecture and Design Principles

A designer is choosing between active-active and active-standby firewall insertion for tenant traffic in an EVPN fabric. Which consideration is most relevant?

  1. Active-active insertion removes the need for VXLAN encapsulation
  2. Active-standby insertion eliminates the need for any routing in the fabric
  3. Firewall insertion has no effect on fabric path design
  4. Active-active insertion typically requires symmetric flow handling and consistent state across both firewalls Correct answer.

Explanation

Active-active firewall insertion demands symmetric forwarding and synchronized session state so both directions of a flow traverse the same device. Active-standby does not remove fabric routing needs, and service insertion directly shapes path design.

Question 26Data Center Architecture and Design Principles

In a spine-leaf IP fabric, what is the defining characteristic of a fully meshed Layer 3 underlay?

  1. Every leaf peers with every other leaf over a full mesh of point-to-point links
  2. Each leaf peers only with every spine, and spines do not peer with each other Correct answer.
  3. Leaves connect to a shared Layer 2 broadcast domain that spans all spines
  4. Spines peer only with other spines to aggregate traffic between pods

Explanation

A standard Layer 3 spine-leaf underlay is built so that each leaf has an adjacency with every spine, while spines remain unconnected to each other. This keeps the topology simple and lets ECMP distribute traffic across all spines. Adding leaf-to-leaf or spine-to-spine links would break the predictable two-tier model.

Question 27Data Center Architecture and Design Principles

Which statement describes the role of a border leaf (also called a border gateway) in a data center fabric?

  1. It provides the connection point between the fabric and external networks such as WAN, Internet, or other data centers Correct answer.
  2. It terminates all server-facing access ports in a rack
  3. It is the only device allowed to run BGP within the underlay
  4. It replaces the spines by performing all inter-leaf forwarding within the fabric

Explanation

A border leaf sits at the edge of the fabric and provides connectivity to networks outside the fabric, including WAN, Internet, and other sites. It typically hosts external BGP sessions and may perform functions such as NAT or firewall handoff. The internal fabric forwarding remains the job of the spines.

Question 28Data Center Architecture and Design Principles

A designer is sizing a fabric where each leaf needs 48 server ports at 25 Gbps and two 100 Gbps uplinks to each spine. Which factor most directly limits how many leaves a single pair of spines can support?

  1. The VXLAN VNI range available on the leaves
  2. The number of MAC addresses learned on each leaf
  3. The number of VLANs configured on the leaves
  4. The spine's total uplink port capacity divided by the per-leaf uplink count Correct answer.

Explanation

The number of leaves a spine pair can support is bounded by how many uplink ports the spines have and how many of those ports each leaf consumes. If each leaf uses two uplinks per spine, the spine port count divided by that per-leaf consumption gives the practical leaf limit. VLAN, MAC, and VNI counts do not constrain physical port fan-out in the same direct way.

Question 29Data Center Architecture and Design Principles

In a Junos EVPN-VXLAN fabric, which routing instance type is used to hold the tenant's Layer 2 and Layer 3 tenant routing information separately from the underlay?

  1. inet.0
  2. virtual-switch or evpn routing instance type Correct answer.
  3. iso.0
  4. inet.3

Explanation

Tenant separation in an EVPN-VXLAN fabric is achieved with a dedicated routing instance, such as a virtual-switch or EVPN instance type, that carries the tenant's bridging and routing state. The underlay remains in the default inet.0 instance. Keeping tenant state in its own instance isolates it from the fabric's own routing.

Question 30Data Center Architecture and Design Principles

A design team wants to reduce the blast radius of a single leaf failure for dual-homed servers. Which design element most directly provides link-level redundancy at the server without relying on server-side bonding?

  1. A multihomed Ethernet segment with EVPN all-active mode Correct answer.
  2. A single uplink from the server to one leaf
  3. A static default route on the server pointing to one leaf
  4. A spanning-tree root bridge on each leaf

Explanation

An EVPN multihomed Ethernet segment lets a server connect to two leaves that act as one logical segment, with all-active forwarding providing redundancy and load sharing. This removes dependence on server-side bonding for basic link redundancy. A single uplink or a static route to one leaf leaves the server vulnerable to that leaf failing.

Question 31Data Center Architecture and Design Principles

Which statement best describes why a routed server access model can simplify fabric operations compared with a Layer 2 access model?

  1. It removes the need for BGP in the underlay
  2. It eliminates the need for any IP addressing on servers
  3. It guarantees that every server can reach every VLAN without routing
  4. It removes large broadcast domains and the associated spanning-tree and flooding behavior from the access layer Correct answer.

Explanation

In a routed server access model each server or host connects to a routed port, so the access layer no longer carries large Layer 2 broadcast domains. That removes spanning tree and much of the flooding behavior that complicates Layer 2 access designs. Servers still need IP addressing, and routing between segments is still required.

Question 32Data Center Architecture and Design Principles

What is the purpose of a distributed anycast gateway in an EVPN-VXLAN fabric?

  1. To assign VNIs automatically to each VLAN
  2. To force all inter-subnet traffic through a single central gateway leaf
  3. To replace the underlay routing protocol with a static default route
  4. To allow every leaf to act as the default gateway for its locally attached hosts using the same IP and MAC Correct answer.

Explanation

A distributed anycast gateway configures the same gateway IP and MAC on every leaf, so each host uses its locally attached leaf as the default gateway. Traffic is then routed at the first hop, avoiding tromboning to a central gateway. This improves efficiency and reduces latency for inter-subnet flows.

Question 33Data Center Architecture and Design Principles

A fabric uses a 3:1 oversubscription ratio at the leaf-to-spine layer. What does this ratio describe?

  1. The ratio of BGP peers to physical interfaces on a spine
  2. The ratio of total downstream host bandwidth to total uplink bandwidth from a leaf Correct answer.
  3. The ratio of VXLAN VNIs to VLANs configured on a leaf
  4. The ratio of spine ports to leaf ports in the fabric

Explanation

Oversubscription compares the aggregate bandwidth available to hosts on the downstream side of a leaf with the aggregate uplink bandwidth toward the spines. A 3:1 ratio means hosts can theoretically demand three times the uplink capacity. Lower ratios favor east-west heavy workloads, while higher ratios reduce cost.

Question 34Data Center Architecture and Design Principles

Which design choice best supports a large Layer 2 domain stretched between two data centers while limiting failure propagation?

  1. Extending a single large VLAN across both sites with spanning tree
  2. Bridging the two sites with a single trunk link between core switches
  3. Using EVPN-VXLAN with a dedicated border leaf pair at each site and controlled route advertisement Correct answer.
  4. Running OSPF between the two sites to carry Layer 2 frames

Explanation

EVPN-VXLAN allows Layer 2 extension over a Layer 3 underlay, and using dedicated border leaf pairs at each site keeps the failure domain controlled. Route advertisement can be scoped so that only intended prefixes and MACs are exchanged. Spanning tree across sites or a single trunk link creates large failure domains and loop risk.

Question 35Data Center Architecture and Design Principles

What does an EVPN type-2 route advertise in an EVPN-VXLAN fabric?

  1. An Ethernet segment identifier for a multihomed device
  2. An IP prefix reachable through a VTEP
  3. A MAC and IP address pair associated with a host reachable through a VTEP Correct answer.
  4. A multicast group membership for a VNI

Explanation

Type-2 routes carry MAC and IP binding information for hosts, allowing remote leaves to learn how to reach a specific host through a given VTEP. They are central to bridging and to ARP suppression. Type-5 routes, by contrast, carry IP prefixes rather than host MAC/IP bindings.

Question 36Data Center Architecture and Design Principles

In a fabric design, why is an MTU of at least 1600 bytes commonly recommended on underlay links carrying VXLAN traffic?

  1. To allow jumbo frames for storage traffic only
  2. To allow more VLAN tags to be stacked on a trunk
  3. To permit BGP to carry larger update messages
  4. To accommodate the VXLAN, UDP, and IP encapsulation overhead without fragmenting tenant frames Correct answer.

Explanation

VXLAN adds encapsulation overhead of roughly 50 bytes, so underlay links must carry the original frame plus that overhead without fragmentation. Setting the underlay MTU to at least 1600 bytes provides headroom for standard 1500-byte tenant frames. Fragmentation or drops would otherwise occur for full-size tenant packets.

Question 37Data Center Architecture and Design Principles

Which statement about the relationship between a VNI and a VLAN in a Junos EVPN-VXLAN leaf is correct?

  1. A VNI is a global identifier mapped to a local VLAN on each leaf, so the same VNI can map to different VLAN IDs on different leaves Correct answer.
  2. A VNI is only used on border leaves and not on regular leaves
  3. A VNI must always equal the VLAN ID on every leaf
  4. A VLAN can map to multiple VNIs on the same leaf simultaneously

Explanation

The VNI is the fabric-wide identifier carried in the VXLAN header, while the VLAN ID is locally significant on each leaf. Mapping lets each leaf use its own VLAN numbering while still participating in the same overlay segment. This decouples local switching configuration from the overlay identity.

Question 38Data Center Architecture and Design Principles

A design must support tenant traffic inspection by a firewall while preserving symmetric forwarding through the fabric. Which approach best meets this requirement?

  1. Inserting a single active-standby firewall pair as a one-armed service on one border leaf only
  2. Using a service leaf with a firewall pair and steering traffic so that both directions traverse the same firewall member Correct answer.
  3. Relying on ECMP hashing alone to send both directions through the same firewall
  4. Placing the firewall inline between the spines and disabling ECMP

Explanation

Symmetric forwarding requires that both directions of a flow traverse the same stateful firewall member. A service leaf design with traffic steering, such as policy-based routing or service chaining, keeps flows symmetric. Relying on ECMP hashing or a single one-armed firewall does not guarantee both directions take the same path.

Question 39Data Center Architecture and Design Principles

In an EVPN-VXLAN fabric, what is the function of an inclusive multicast route (type-3)?

  1. It advertises IP prefixes for inter-subnet routing
  2. It elects a designated forwarder for a multihomed Ethernet segment
  3. It advertises host MAC and IP bindings to remote leaves
  4. It advertises the VTEPs that participate in a given VNI so that broadcast, unknown unicast, and multicast traffic can be delivered to them Correct answer.

Explanation

Type-3 inclusive multicast routes tell other VTEPs which endpoints are members of a VNI, building the broadcast, unknown unicast, and multicast delivery tree for that VNI. They are fundamental to how overlay flooding is scoped. Host bindings and prefix routes are carried by other route types.

Question 40Data Center Architecture and Design Principles

Which consideration most directly drives the decision to deploy a super-spine layer in a large data center fabric?

  1. The need to reduce the number of VLANs per leaf
  2. The need to scale beyond the port capacity of a single spine tier while keeping the fabric non-blocking Correct answer.
  3. The need to eliminate BGP from the underlay
  4. The need to centralize all default gateways on one device

Explanation

A super-spine layer is added when the fabric must grow beyond what a single spine tier can support in ports and bandwidth. It provides another tier of aggregation so that pods can interconnect while maintaining non-blocking or near-non-blocking behavior. VLAN counts, routing protocol choice, and gateway placement are separate design decisions.

Question 41Data Center Switching and Fabric Technologies

In a VXLAN EVPN fabric, which control-plane protocol is used to advertise MAC addresses and IP addresses between leaf switches?

  1. STP with PVST+ enhancements
  2. PIM-SM with anycast RPs
  3. OSPFv2 with type-2 LSAs
  4. MP-BGP with the EVPN address family Correct answer.

Explanation

VXLAN EVPN fabrics use MP-BGP with the EVPN address family to distribute MAC and IP reachability information as route types 2 and 5. This replaces data-plane MAC learning over the overlay with a control-plane-driven model. OSPF, PIM, and STP do not carry EVPN route types.

Question 42Data Center Switching and Fabric Technologies

Which routing protocol is most commonly deployed as the underlay IGP in a VXLAN EVPN data center fabric?

  1. BGP with full mesh iBGP
  2. IS-IS or OSPF with ECMP Correct answer.
  3. EIGRP stub routing
  4. RIP version 2

Explanation

Modern VXLAN EVPN fabrics typically use a link-state IGP such as OSPF or IS-IS in the underlay so every leaf and spine has equal-cost paths to all loopbacks. This enables ECMP load sharing across the spine layer. RIP and EIGRP stub designs do not scale to this role in typical data center fabrics.

Question 43Data Center Switching and Fabric Technologies

What is the primary purpose of a vPC (virtual PortChannel) domain in a Cisco Nexus data center design?

  1. To encrypt traffic between spine and leaf switches
  2. To provide Layer 3 routing between VLANs on a single switch
  3. To replace BGP with a proprietary routing protocol
  4. To allow two switches to appear as one logical device to downstream devices Correct answer.

Explanation

A vPC domain lets two Nexus switches present a single logical PortChannel endpoint to attached servers or switches, providing active-active forwarding and redundancy. It does not itself perform inter-VLAN routing or encryption. It also does not replace routing protocols.

Question 44Data Center Switching and Fabric Technologies

Which VXLAN component is responsible for encapsulating original Ethernet frames into UDP packets?

  1. The VTEP Correct answer.
  2. The anycast gateway
  3. The PIM rendezvous point
  4. The vPC peer-link

Explanation

A VTEP (VXLAN Tunnel Endpoint) performs the encapsulation and decapsulation of tenant Ethernet frames into and out of VXLAN UDP packets. It maps the original frame to a VNI and adds the outer IP and UDP headers. The peer-link, anycast gateway, and PIM RP serve different functions.

Question 45Data Center Switching and Fabric Technologies

In a spine-leaf fabric, what is the primary reason every leaf connects to every spine?

  1. To provide predictable any-to-any ECMP forwarding with consistent latency Correct answer.
  2. To guarantee a single active path between any two hosts
  3. To allow spanning tree to block redundant links
  4. To eliminate the need for any routing protocol

Explanation

A full mesh between leaves and spines gives every leaf an equal-cost path to every other leaf through any spine, enabling ECMP and predictable hop count and latency. It does not remove the need for routing protocols, nor does it rely on STP blocking. Redundancy comes from multiple equal paths rather than a single active path.

Question 46Data Center Switching and Fabric Technologies

Which Cisco Nexus feature provides a distributed anycast gateway so that a host can use the same default gateway IP regardless of which leaf it attaches to?

  1. VRRP with priority tuning
  2. Distributed anycast gateway (AGW) Correct answer.
  3. HSRP with preemption
  4. Proxy ARP on the spine

Explanation

The distributed anycast gateway configures the same gateway IP and MAC on every leaf, so a host's first-hop gateway is always local to its attached leaf. This enables optimal forwarding and seamless VM mobility. HSRP and VRRP provide a single active gateway per subnet, and proxy ARP does not create a distributed gateway.

Question 47Data Center Switching and Fabric Technologies

Which protocol does a VXLAN EVPN fabric use for underlay multicast replication when ingress replication is not used?

  1. IGMP snooping only
  2. OSPF multicast extensions
  3. PIM-SM or PIM-Bidir Correct answer.
  4. BGP MVPN only

Explanation

When the underlay uses multicast replication for BUM traffic, PIM-SM or PIM-Bidir builds the multicast distribution trees in the underlay. IGMP snooping alone does not build routed multicast trees across the fabric. OSPF and BGP MVPN do not serve as the underlay multicast routing protocol in this design.

Question 48Data Center Switching and Fabric Technologies

What does the VXLAN VNI identify within an encapsulated packet?

  1. The vPC domain identifier
  2. The physical spine uplink
  3. The tenant Layer 2 or Layer 3 segment Correct answer.
  4. The BGP autonomous system number

Explanation

The 24-bit VNI in the VXLAN header identifies the tenant segment, mapping to a VLAN or bridge domain on the VTEP. It allows up to roughly 16 million segments to be isolated over a shared IP underlay. Spine uplinks, AS numbers, and vPC domain IDs are unrelated to the VNI.

Question 49Data Center Switching and Fabric Technologies

A leaf switch in a VXLAN EVPN fabric must forward a broadcast frame from a host. Which mechanism does it use when the fabric is configured for ingress replication?

  1. It floods the frame out all spine uplinks as native Ethernet
  2. It sends the frame to the multicast group of the bridge domain
  3. It replicates the frame as unicast VXLAN packets to every remote VTEP that has the VNI Correct answer.
  4. It drops the frame until a receiver is learned

Explanation

With ingress replication, the ingress VTEP makes a copy of the BUM frame for each remote VTEP that participates in the VNI and sends them as unicast VXLAN packets. Multicast group flooding applies to multicast-replication designs. Native flooding over spines and dropping BUM traffic are not correct behaviors.

Question 50Data Center Switching and Fabric Technologies

An engineer wants a server to use active-active dual-homed connections to two different leaf switches that are not in the same vPC domain. Which approach is appropriate?

  1. Use a vPC or EVPN multihoming solution that presents a single logical link Correct answer.
  2. Rely on STP to block one server uplink
  3. Configure a standard PortChannel on the server to both leaves
  4. Connect the server to only one leaf for simplicity

Explanation

Dual-homing a server to two independent leaves requires a multichassis solution such as vPC or EVPN multihoming so the two links appear as one logical bundle. A standard PortChannel across two independent switches is not supported and would be blocked or unstable. STP blocking would leave one uplink unused, and single-homing sacrifices redundancy.