I85PREP
⌘ K
DCNP study materials
DCNP study resources

DCNP practice questions

One question at a time. Build your DCNP 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 networking fundamentals and architectures

In a data center, which cabling media is specified for 10GBASE-T links up to 100 meters?

  1. Category 5e twisted pair
  2. OM3 multimode fiber
  3. Single-mode fiber
  4. Category 6A twisted pair Correct answer.

Explanation

10GBASE-T over the full 100 m channel requires cabling with extended bandwidth and improved alien crosstalk performance, which Category 6A provides. Category 6 is limited to shorter distances (about 55 m) in practice, and fiber media are not used for BASE-T copper ports.

Question 2Data center networking fundamentals and architectures

Which topology is characteristic of a traditional three-tier data center network design?

  1. Full mesh between every server
  2. Core, aggregation, and access layers Correct answer.
  3. Spine and leaf layers only
  4. Ring with dual counter-rotating paths

Explanation

The three-tier (or three-layer) design separates core, aggregation (distribution), and access functions, each with a distinct role in forwarding and policy. Spine-leaf is a different two-tier architecture, and rings or full server meshes are not the standard three-tier model.

Question 3Data center networking fundamentals and architectures

What is the primary purpose of a spine-leaf (Clos) fabric in a data center?

  1. To replace all copper cabling with passive optical splitters
  2. To concentrate all inter-server traffic through a single core switch
  3. To eliminate the need for Layer 3 routing in the data center
  4. To provide predictable any-to-any server-to-server latency and bandwidth Correct answer.

Explanation

A spine-leaf fabric gives every leaf an equal-cost path to every other leaf through the spines, producing consistent hop count and balanced bandwidth. This predictable east-west performance is the main reason the design replaced three-tier aggregation for server traffic.

Question 4Data center networking fundamentals and architectures

Which protocol provides a loop-free Layer 2 topology by computing a single active path tree in a data center access layer?

  1. OSPF
  2. BGP
  3. Spanning Tree Protocol Correct answer.
  4. LACP

Explanation

Spanning Tree Protocol blocks redundant links to create a loop-free logical tree in a Layer 2 domain. OSPF and BGP are Layer 3 routing protocols, and LACP bundles links rather than preventing Layer 2 loops.

Question 5Data center networking fundamentals and architectures

In a data center, what does east-west traffic primarily describe?

  1. Traffic between users on the internet and the data center
  2. Traffic between servers within the data center Correct answer.
  3. Traffic between the data center and a public cloud region
  4. Traffic between the data center and a remote backup site

Explanation

East-west traffic refers to server-to-server communication inside the data center, which dominates in distributed and virtualized applications. North-south traffic, by contrast, crosses the data center boundary toward users or other sites.

Question 6Data center networking fundamentals and architectures

Which device function is performed by a top-of-rack (ToR) switch?

  1. Performing application load balancing across web servers
  2. Providing storage controller RAID functions
  3. Terminating WAN circuits from service providers
  4. Aggregating server connections within a single rack and uplinking to the fabric Correct answer.

Explanation

A top-of-rack switch connects the servers in one rack and provides uplinks into the aggregation or spine layer. WAN termination, RAID control, and application load balancing are functions of other devices such as routers, storage controllers, and load balancers.

Question 7Data center networking fundamentals and architectures

What does a /24 prefix length indicate in an IPv4 data center subnet?

  1. A subnet supporting only 24 hosts
  2. A 24-bit network mask with 254 usable host addresses Correct answer.
  3. A network using 24 VLANs
  4. A 24-bit host portion with 16 million addresses

Explanation

A /24 mask uses 24 network bits and 8 host bits, giving 256 addresses of which 254 are usable after reserving the network and broadcast addresses. Prefix length counts network bits, not hosts or VLANs.

Question 8Data center networking fundamentals and architectures

What is the role of a VLAN in a data center switching design?

  1. To assign a unique physical cable to each server
  2. To encrypt all traffic between switches
  3. To guarantee bandwidth for storage traffic
  4. To logically segment a Layer 2 domain independent of physical switch location Correct answer.

Explanation

A VLAN creates a logical broadcast domain that can span multiple switches, allowing segmentation without regard to physical rack layout. It does not encrypt traffic, reserve bandwidth, or map to individual cables.

Question 9Data center networking fundamentals and architectures

A leaf switch has four equal-cost uplinks to four spine switches. Which mechanism distributes flows across all four links simultaneously?

  1. Spanning Tree Protocol
  2. VRRP
  3. Equal-cost multipath (ECMP) Correct answer.
  4. LACP

Explanation

ECMP installs multiple equal-cost next hops and hashes flows across them, so all four uplinks carry traffic at once. Spanning Tree would block redundant paths, VRRP is a gateway redundancy protocol, and LACP aggregates links to a single peer rather than across separate spines.

Question 10Data center networking fundamentals and architectures

In a spine-leaf fabric, why is Layer 3 routing typically extended to the leaf switches?

  1. To avoid assigning IP addresses to servers
  2. To allow a single subnet to span the entire fabric
  3. To remove the need for VLANs on access ports
  4. To enable ECMP and contain broadcast domains at each leaf Correct answer.

Explanation

Routing at the leaf lets each leaf terminate its own subnets, use ECMP across all spines, and keep Layer 2 broadcast domains small. Servers still need IP addresses, and a single stretched subnet would enlarge failure domains rather than contain them.

Question 11Data center networking fundamentals and architectures

Which overlay technology encapsulates Layer 2 frames inside IP packets to extend segments across a Layer 3 data center fabric?

  1. VXLAN Correct answer.
  2. OSPFv2
  3. STP
  4. SNMP

Explanation

VXLAN uses MAC-in-UDP encapsulation to carry Layer 2 frames over a routed IP underlay, enabling large-scale virtual segments. STP and OSPF are control protocols, and SNMP is used for management.

Question 12Data center networking fundamentals and architectures

What is the function of a VXLAN tunnel endpoint (VTEP)?

  1. To assign DHCP addresses to physical servers
  2. To act as the default gateway for the management network
  3. To encapsulate and decapsulate tenant frames between the overlay and the underlay Correct answer.
  4. To terminate storage-area network sessions

Explanation

A VTEP sits at the edge of the overlay and performs encapsulation on ingress and decapsulation on egress, mapping VNIs to underlay endpoints. DHCP, SAN termination, and management gateway functions belong to other services.

Question 13Data center networking fundamentals and architectures

Which IEEE standard defines link aggregation for combining multiple physical links into one logical channel?

  1. IEEE 802.3af
  2. IEEE 802.1X
  3. IEEE 802.3ad Correct answer.
  4. IEEE 802.1Q

Explanation

IEEE 802.3ad defines link aggregation, later refined by 802.1AX, which presents multiple links as one logical interface. 802.1Q covers VLAN tagging, 802.1X covers port authentication, and 802.3af covers Power over Ethernet.

Question 14Data center networking fundamentals and architectures

In a data center design, what is the main benefit of using a 40/100 GbE uplink from leaf to spine rather than 10 GbE?

  1. It reduces the need for Layer 3 routing
  2. It eliminates the need for ECMP
  3. It allows a single spine to serve all leaves without redundancy
  4. It provides higher oversubscription headroom for east-west traffic Correct answer.

Explanation

Higher-capacity uplinks reduce the oversubscription ratio between server-facing and fabric-facing ports, which matters because east-west traffic dominates. They do not remove the need for routing, ECMP, or redundant spines.

Question 15Data center networking fundamentals and architectures

A data center uses a 1:4 oversubscription ratio at the leaf. What does this ratio mean?

  1. Four times as much server-facing bandwidth as uplink bandwidth Correct answer.
  2. Four spines per leaf
  3. Four uplinks per server
  4. Four times as much uplink bandwidth as server-facing bandwidth

Explanation

Oversubscription compares access-side capacity to uplink capacity; 1:4 means the server-facing side can offer four times the bandwidth of the uplinks. This is a capacity ratio, not a count of uplinks or spines.

Question 16Data center networking fundamentals and architectures

Which protocol is commonly used as a routing protocol in a large spine-leaf underlay to support ECMP and fast convergence?

  1. BGP Correct answer.
  2. RIP
  3. IGMP
  4. LLDP

Explanation

eBGP is widely deployed as the underlay routing protocol in large fabrics because it scales well and supports multipath. RIP converges slowly and scales poorly, IGMP manages multicast membership, and LLDP is a neighbor discovery protocol.

Question 17Data center networking fundamentals and architectures

What is the purpose of a first-hop redundancy protocol such as VRRP in a data center?

  1. To authenticate servers before they join the network
  2. To balance Layer 2 frames across multiple uplinks
  3. To provide a shared virtual default gateway address across redundant routers Correct answer.
  4. To prevent Layer 2 loops in the access layer

Explanation

VRRP lets multiple routers share a virtual IP and MAC so hosts keep a consistent default gateway if one router fails. Link balancing, loop prevention, and authentication are handled by other protocols.

Question 18Data center networking fundamentals and architectures

Which cabling practice best supports future upgrades from 10G to 25G or 40G in a data center spine-leaf build?

  1. Using single-pair copper for all spine uplinks
  2. Deploying structured fiber trunks with MPO connectors between rows Correct answer.
  3. Installing Category 5e patch cords for all server links
  4. Relying on wireless links between racks

Explanation

Pre-terminated MPO fiber trunks provide high strand counts and support parallel optics for 40G and 100G, easing later upgrades. Category 5e and single-pair copper cannot carry these rates over typical distances, and wireless is not used for fabric links.

Question 19Data center networking fundamentals and architectures

In a spine-leaf fabric, why are server subnets typically not stretched across multiple leaf switches?

  1. To eliminate the need for a routing protocol
  2. To reduce the number of available IP addresses
  3. To avoid using VLANs on access ports
  4. To keep broadcast and failure domains small and preserve predictable routing Correct answer.

Explanation

Keeping subnets local to a leaf limits broadcast scope and failure impact while allowing ECMP routing between leaves. Stretching subnets enlarges Layer 2 domains and complicates routing, and it does not reduce available addresses or remove the need for routing.

Question 20Data center networking fundamentals and architectures

A design team must connect 48 servers per rack with dual-homed 25G links and provide at least 1:1 non-blocking east-west capacity. Which approach meets this requirement?

  1. One ToR leaf switch with a single 40G uplink to one spine
  2. Two ToR switches with 10G uplinks to two spines
  3. One ToR switch with 1G uplinks to all spines
  4. Two ToR leaf switches with MLAG and 100G uplinks to every spine Correct answer.

Explanation

Dual-homed servers need two ToR switches with MLAG for link redundancy, and 100G uplinks to every spine deliver enough fabric capacity to avoid oversubscription. Single-homed or low-speed uplink options either lose redundancy or create heavy oversubscription.

Question 21Data center networking fundamentals and architectures

An operator observes that one ECMP uplink from a leaf is saturated while the other three are lightly used. Which cause is most consistent with this symptom?

  1. Flow hashing is polarized because a few large flows share the same hash Correct answer.
  2. VRRP has failed over to a standby router
  3. The spines are running different routing protocols
  4. Spanning Tree has blocked the other three uplinks

Explanation

ECMP balances per flow, not per packet, so a small number of high-bandwidth flows can hash to the same uplink and leave others idle. Spanning Tree blocking, VRRP failover, and protocol mismatch would not produce this specific uneven pattern while all links remain active.

Question 22Data center networking fundamentals and architectures

A VXLAN overlay is being designed over an IP underlay. Which underlay property is most critical for overlay performance?

  1. Running STP on every leaf
  2. MTU sized to accommodate the extra encapsulation header Correct answer.
  3. Use of Category 5e cabling throughout
  4. Disabling ECMP on all spines

Explanation

VXLAN adds encapsulation overhead, so the underlay must support a larger MTU to avoid fragmentation of encapsulated frames. Cabling category, ECMP, and STP are not the decisive factor for overlay packet handling.

Question 23Data center networking fundamentals and architectures

A fabric uses eBGP between leaves and spines with unique AS numbers per leaf. What is the main reason for assigning a distinct ASN to each leaf?

  1. To avoid the need for ECMP
  2. To prevent a leaf from accepting routes that loop back through another leaf Correct answer.
  3. To reduce the number of IP addresses used on uplinks
  4. To allow the spines to run VRRP

Explanation

Unique leaf ASNs let BGP path selection detect and reject routes that would loop through another leaf, which is important in large fabrics. Address conservation, VRRP, and ECMP are unrelated to ASN assignment.

Question 24Data center networking fundamentals and architectures

A data center must support lossless Ethernet for storage traffic while also carrying ordinary IP traffic on the same links. Which mechanism is designed for this requirement?

  1. DHCP snooping
  2. Global pause frames for the entire link
  3. Priority-based flow control with pause per traffic class Correct answer.
  4. Spanning Tree root guard

Explanation

Priority-based flow control pauses only a specific traffic class, allowing lossless treatment for storage while other classes continue. Global pause halts the whole link, and root guard or DHCP snooping address unrelated Layer 2 security concerns.

Question 25Data center networking fundamentals and architectures

Which statement accurately compares north-south and east-west traffic in a modern data center?

  1. East-west traffic refers only to traffic leaving the data center
  2. East-west traffic typically exceeds north-south traffic because applications are distributed across many servers Correct answer.
  3. North-south traffic stays within a single rack
  4. North-south traffic always exceeds east-west traffic in virtualized environments

Explanation

Distributed and microservice applications generate heavy server-to-server communication, so east-west traffic usually dominates. North-south traffic crosses the boundary to users or other sites, and it is not confined to a rack.

Question 26Data center networking fundamentals and architectures

In a three-tier data center topology, which layer aggregates access switches and enforces policy before traffic reaches the core?

  1. The core layer
  2. The distribution (aggregation) layer Correct answer.
  3. The storage layer
  4. The access layer

Explanation

The three-tier model separates access, distribution, and core functions. Access switches connect servers, the core provides high-speed transport between distribution blocks, and the distribution layer aggregates access switches and applies routing and policy boundaries.

Question 27Data center networking fundamentals and architectures

Which statement describes a cross-connect in a data center physical infrastructure?

  1. A physical patching point that links two ports, such as a server NIC to a switch port or a switch port to a patch panel Correct answer.
  2. A power distribution unit feeding redundant supplies
  3. A logical tunnel that carries VLAN tags between sites
  4. A routing adjacency formed between two spine switches

Explanation

A cross-connect is a physical cabling relationship between two endpoints, usually made at a patch panel or fiber distribution panel. It is documented in cable management records so that moves, adds, and changes can be traced without disturbing other connections.

Question 28Data center networking fundamentals and architectures

What is the primary purpose of structured cabling standards such as TIA-942 in a data center build?

  1. To define a consistent, documented cabling hierarchy and media performance so that systems interoperate and can be upgraded predictably Correct answer.
  2. To mandate a specific switch vendor for spine-leaf fabrics
  3. To set the maximum number of VLANs allowed on a trunk
  4. To specify the routing protocol used between leaf and spine switches

Explanation

Structured cabling standards describe topology, media, connector, and documentation practices rather than vendor or protocol choices. Following them gives predictable channel performance and makes later speed upgrades and troubleshooting easier.

Question 29Data center networking fundamentals and architectures

Which traffic pattern do modern data center fabrics primarily need to optimize for?

  1. North-south client-to-server flows entering through the WAN edge
  2. Management-plane polling from a central monitoring server
  3. Backup traffic scheduled during off-peak hours
  4. East-west server-to-server flows generated by distributed applications Correct answer.

Explanation

Distributed applications, clustering, and storage replication generate large volumes of server-to-server traffic inside the fabric. Designs such as spine-leaf with equal-cost multipathing exist mainly to carry that east-west load with predictable latency and high bisection bandwidth.

Question 30Data center networking fundamentals and architectures

A fabric uses a two-tier spine-leaf design where every leaf connects to every spine. What property does this full-mesh guarantee?

  1. Any leaf can reach any other leaf with a consistent number of fabric hops Correct answer.
  2. Bandwidth between any two leaves is limited to a single 10G link
  3. Leaf-to-leaf traffic must always pass through a dedicated services chassis
  4. Spine switches must run different routing protocols from leaf switches

Explanation

In a full-mesh spine-leaf fabric each leaf is exactly one hop from every spine, so leaf-to-leaf paths are always leaf-spine-leaf. This uniform hop count keeps latency predictable and lets equal-cost multipathing spread flows evenly.

Question 31Data center networking fundamentals and architectures

Why is a Clos-based spine-leaf fabric considered scalable compared with a traditional three-tier design?

  1. It eliminates the need for any routing protocol in the fabric
  2. It removes the requirement for redundant power in network racks
  3. Adding spine switches increases aggregate bisection bandwidth without requiring a redesign of existing leaf uplinks Correct answer.
  4. It allows a single access switch to serve every rack in the data center

Explanation

Clos fabrics scale horizontally: each new spine adds another parallel path that every leaf can use, raising total fabric capacity. Because the leaf-to-spine cabling pattern is uniform, growth is incremental rather than requiring a new hierarchical tier.

Question 32Data center networking fundamentals and architectures

Which statement best describes the role of a border leaf in a spine-leaf fabric?

  1. It terminates all storage-area network traffic from servers
  2. It acts as the time synchronization source for the fabric
  3. It stores the routing table for every other leaf in the fabric
  4. It provides connectivity between the fabric and external networks such as WAN, internet, or legacy data center domains Correct answer.

Explanation

A border leaf is a leaf switch positioned to connect the fabric to outside networks, hosting external routing peers and security boundaries. Ordinary leaf switches focus on server attachment, so separating the border role keeps external policy and route scale off the server-facing leaves.

Question 33Data center networking fundamentals and architectures

Which technology allows a single physical data center switch to be managed and forwarded as multiple independent logical switches?

  1. VLAN trunking
  2. Jumbo frame support
  3. Network virtualization through switch partitioning Correct answer.
  4. Link aggregation

Explanation

Switch partitioning, sometimes called logical switch or VDC-style virtualization, divides one physical chassis into separate logical devices with their own configuration and forwarding contexts. VLAN trunking and link aggregation operate within a single logical switch rather than creating independent ones.

Question 34Data center networking fundamentals and architectures

A design must support lossless transport for storage traffic and ordinary IP traffic on shared Ethernet links. Which capability is required?

  1. Priority-based flow control with pause per traffic class Correct answer.
  2. Forcing all storage traffic through the border leaf
  3. A single flat VLAN for all traffic types
  4. Disabling equal-cost multipathing on storage-facing leaves

Explanation

Lossless Ethernet uses priority-based flow control so that a pause frame affects only the class of service carrying storage, leaving ordinary IP traffic unthrottled. This per-priority pause behavior is what allows converged links to carry both traffic types safely.

Question 35Data center networking fundamentals and architectures

Which statement accurately distinguishes north-south from east-west traffic in a data center?

  1. East-west traffic always traverses the internet edge first
  2. North-south crosses the data center boundary, while east-west stays within the fabric between internal endpoints Correct answer.
  3. North-south stays within the fabric, while east-west crosses the WAN edge
  4. Both terms describe the same flow measured at different times

Explanation

North-south traffic enters or leaves the data center through edge devices, typically between clients and servers. East-west traffic moves between servers, storage, and appliances inside the fabric, and it dominates volume in distributed application environments.

Question 36Data center networking fundamentals and architectures

An operator needs a fabric that supports large Layer 2 domains across racks without stretching spanning tree across the whole data center. Which approach fits?

  1. An IP-routed underlay with a VXLAN overlay providing Layer 2 segments at the edge Correct answer.
  2. Spanning tree with a reduced bridge priority on all switches
  3. Replacing all leaf-spine links with hubs
  4. A single large access VLAN trunked to every rack switch

Explanation

Routing the underlay and carrying Layer 2 segments in a VXLAN overlay confines broadcast domains to the overlay edge and avoids a large spanning tree topology. The underlay uses routed equal-cost paths, while the overlay delivers the Layer 2 stretch applications require.

Question 37Data center networking fundamentals and architectures

A leaf has four equal-cost uplinks to four spines. An operator wants per-flow hashing to use all four links. Which condition must hold?

  1. The leaf must disable routing and rely on bridging
  2. The underlay must install all four paths in the forwarding table with equal cost Correct answer.
  3. Each uplink must be in a different VLAN
  4. The spines must have different hostnames

Explanation

Equal-cost multipathing requires the routing table to contain multiple next hops of identical metric so the forwarding engine can hash flows across them. Without equal-cost paths installed, traffic collapses onto a single uplink regardless of hashing configuration.

Question 38Data center networking fundamentals and architectures

A fabric is designed for a 3:1 oversubscription ratio at the leaf tier. What does this ratio express?

  1. The ratio of power capacity to cooling capacity in the rack
  2. The ratio of aggregate downlink bandwidth to servers versus aggregate uplink bandwidth toward the spine Correct answer.
  3. The ratio of Layer 3 prefixes to Layer 2 MAC addresses
  4. The ratio of spine switches to leaf switches in the fabric

Explanation

Oversubscription compares the bandwidth available to attached servers with the bandwidth available toward the rest of the fabric. A 3:1 ratio means servers can collectively demand three times the uplink capacity, so contention is possible during peak east-west bursts.

Question 39Data center networking fundamentals and architectures

A design requires that a rack keep serving traffic if one leaf switch fails, with servers dual-homed. Which mechanism provides active-active server attachment across two leaves?

  1. Multi-chassis link aggregation between the two leaf switches Correct answer.
  2. Static routes on the server pointing to one leaf
  3. A single uplink from each server to one leaf
  4. Spanning tree blocking one server link permanently

Explanation

Multi-chassis link aggregation lets a server form one logical bond across two separate leaf switches, so both links forward traffic and either leaf can fail without losing connectivity. Spanning tree would instead block one link and leave it idle.

Question 40Data center networking fundamentals and architectures

An operator must connect two data center fabrics from different vendors that both use VXLAN. Which approach most reliably preserves Layer 2 segments across the boundary?

  1. Trunk all VLANs between the fabrics without any mapping
  2. Use a standards-based VXLAN gateway function at the border that maps VNI to VLAN and forwards between the fabrics Correct answer.
  3. Rely on each vendor's proprietary control plane to synchronize automatically
  4. Merge the two underlays into one routing domain with no boundary device

Explanation

Vendor control planes for VXLAN differ, so interoperability is achieved at a gateway that maps VNIs to VLANs and forwards frames between the two overlay domains. This keeps each fabric's control plane independent while preserving the required Layer 2 connectivity.

Question 41Switching, routing, and overlay technologies

In a VXLAN overlay, what does the VNI in the outer header identify?

  1. A single tenant's Layer 2 segment, mapped to a bridge domain Correct answer.
  2. The physical uplink port used by the VTEP
  3. The QoS priority of the encapsulated frame
  4. The BGP autonomous system of the underlay

Explanation

VXLAN uses a 24-bit VNI in the outer header to identify an individual Layer 2 segment, allowing many isolated segments to share one IP underlay. Each VNI is bound to a bridge domain on the VTEP, and the inner Ethernet frame is preserved unchanged.

Question 42Switching, routing, and overlay technologies

Which statement describes the role of a VTEP in an EVPN-VXLAN fabric?

  1. It assigns IP addresses to leaf switches via DHCP
  2. It authenticates users before they reach the overlay
  3. It originates and terminates VXLAN tunnels and maps local VLANs to VNIs Correct answer.
  4. It converts EVPN routes into OSPF link-state advertisements

Explanation

A VTEP is the tunnel endpoint: it encapsulates frames from a local VLAN into VXLAN with the matching VNI and decapsulates traffic arriving from remote VTEPs. It sits at the edge of the overlay, typically on leaf switches.

Question 43Switching, routing, and overlay technologies

Which EVPN route type is used to advertise MAC and IP bindings for hosts learned on a VTEP?

  1. Type 2 Correct answer.
  2. Type 5
  3. Type 1
  4. Type 3

Explanation

EVPN Type 2 (MAC/IP advertisement) routes carry the MAC address and, optionally, the IP address of a host along with the advertising VTEP. They enable both Layer 2 forwarding and ARP/ND suppression in the overlay.

Question 44Switching, routing, and overlay technologies

Which EVPN route type builds the multicast tunnel list used to flood broadcast, unknown unicast, and multicast traffic?

  1. Type 3 Correct answer.
  2. Type 2
  3. Type 4
  4. Type 5

Explanation

The Type 3 inclusive multicast Ethernet tag route tells remote VTEPs which VTEPs belong to the flooding set for a VNI. Ingress replication or a multicast underlay then distributes BUM traffic to that set.

Question 45Switching, routing, and overlay technologies

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

  1. To load-balance traffic across multiple underlay spines
  2. To translate VXLAN VNIs into VLAN IDs at the border
  3. To provide a backup route when BGP peering fails
  4. To let every leaf use the same gateway MAC and IP for a subnet so hosts keep one default gateway Correct answer.

Explanation

With a distributed anycast gateway, each leaf that hosts a subnet uses the same virtual MAC and gateway IP. A host can move between leaves without changing its default gateway or re-ARPing, which keeps first-hop forwarding local.

Question 46Switching, routing, and overlay technologies

Which underlay routing design is most common for a large EVPN-VXLAN spine-leaf fabric?

  1. eBGP between each leaf and each spine with ECMP Correct answer.
  2. IS-IS with a single Level 1 area
  3. Static routes from each leaf to a single spine
  4. OSPF with a single area across all leaves

Explanation

eBGP on each leaf-spine link gives multipath load sharing and limits failure scope, which scales well as the fabric grows. The overlay EVPN address family then runs over these BGP sessions or over a separate overlay peering.

Question 47Switching, routing, and overlay technologies

In a Multichassis LAG (MC-LAG) dual-homing design, what does the Inter-Chassis Control Protocol (ICCP) carry between the peer devices?

  1. Spanning-tree BPDUs for the access VLAN
  2. User data traffic between the two peers
  3. VXLAN encapsulated tenant frames
  4. State and consistency information for the shared LAG Correct answer.

Explanation

ICCP runs over the peer link and synchronizes LAG state and configuration consistency between the two MC-LAG peers. It does not carry user data; the peer link is for control and, in some designs, failover traffic.

Question 48Switching, routing, and overlay technologies

Which protocol provides loop-free Layer 2 multipathing in a data center fabric without relying on spanning tree?

  1. PVST+
  2. RSTP
  3. STP
  4. TRILL Correct answer.

Explanation

TRILL uses IS-IS-based routing at Layer 2 to compute shortest paths and support equal-cost multipathing, avoiding blocked links. The spanning-tree variants all prevent loops by blocking redundant ports.

Question 49Switching, routing, and overlay technologies

A leaf switch must forward a frame to a host on a remote leaf, but the destination MAC is not yet in its EVPN table. How does an EVPN-VXLAN fabric handle this unknown unicast frame?

  1. It is flooded to the VTEPs in the VNI's inclusive multicast set Correct answer.
  2. It is sent only to the spine that owns the destination subnet
  3. It is dropped until the host sends a gratuitous ARP
  4. It is converted to a Type 5 route and routed by the border

Explanation

Until a Type 2 route is learned, unknown unicast is treated as BUM traffic and replicated to the VTEPs listed in the Type 3 inclusive multicast route for that VNI. Once the MAC/IP binding is advertised, subsequent frames are unicast directly to the owning VTEP.

Question 50Switching, routing, and overlay technologies

What is the primary benefit of ARP suppression on a VTEP in an EVPN-VXLAN fabric?

  1. ARP replies are encrypted before crossing the underlay
  2. ARP requests for known hosts are answered locally from the EVPN table, reducing flooding Correct answer.
  3. ARP is replaced entirely by BGP route exchange
  4. ARP requests are redirected to the border leaf for resolution

Explanation

When a VTEP holds a Type 2 route with an IP for the target, it can answer the ARP request itself instead of flooding it across the overlay. This cuts broadcast propagation and speeds up host-to-host resolution.