Skip to content
Tech Interview Prep
Technical interview guide

Routing & Switching Fundamentals

How traffic gets forwarded within a local network (switching) versus across different networks (routing).

Read
45 min
Practice MCQs
25
Interview Q&A
25
Edition
v6
Editorial status
Reviewed
Relevant for
Network Engineer

Scope: Current IETF router, CIDR, OSPFv2/v3, BGP, eBGP policy, VRRP, BFD, EVPN, VXLAN, Cisco Layer 2/STP, and Juniper BGP references reviewed 2026-09-04.

Curated: · Written: · Reviewed:

25 questions

MCQ-1A switch boots with an empty MAC address table. Host A on port 3 sends a frame to Host B, which the switch has never seen, so the switch floods the frame out every other port. Which table entry does the switch create from this frame?
MCQ-2A single access switch carries VLAN 30 on 40 access ports, each with one host, and VLAN 40 on 20 other access ports. The MAC address table ages dynamic entries after 300 seconds. Workstation W (MAC 00:1a:2b:3c:4d:5e) in VLAN 30 was last heard from 6 minutes ago. Server S in VLAN 30 sends one unicast frame to W. How many ports receive a copy of that frame?
MCQ-3A switch port connects to a single server that sends only untagged frames. The port must place those frames in VLAN 20, and any frame arriving tagged for another VLAN must be dropped. Which port configuration does this?
MCQ-6A team bundles four 10 GbE links between two switches with LACP. All four links are up at the same speed and in the same VLAN. Which information does the LACP exchange use to decide whether a given link is added to the aggregate?
MCQ-4SW1 and SW2 are joined by a single 802.1Q trunk on Gi1/0/1, with native VLAN 1 on both ends. SW1 is configured with `switchport trunk allowed vlan 1,10-20`; SW2 is configured with `switchport trunk allowed vlan 1,15-25`. A host in VLAN 12 on SW1 sends a unicast frame to a host in VLAN 12 on SW2. What happens to the frame?
MCQ-5SW1 and SW2 are access switches with two 10 GbE links between them in VLAN 10. STP is disabled. A user connects an unmanaged switch to wall ports on both SW1 and SW2, creating a loop. Within 5 seconds, broadcast traffic in VLAN 10 saturates both 10 GbE links, both switch CPUs hit 100%, and the MAC table entry for one host alternates between Te1/0/1 and Te1/0/2. Which action will restore stable connectivity while keeping both inter-switch links available for future use?
MCQ-7A router's IPv4 forwarding table holds these entries, listed in this order: 10.1.0.0/16 via 192.0.2.3 (metric 10); 10.0.0.0/8 via 192.0.2.2 (metric 5); 10.1.2.0/24 via 192.0.2.4 (metric 30); 10.1.2.128/25 via 192.0.2.5 (metric 20); 0.0.0.0/0 via 192.0.2.1 (metric 1). A packet arrives with destination address 10.1.2.3. Which next hop does the router forward it to?
MCQ-25A router has a static route to 10.20.0.0/16 with a reachable next hop and an OSPF route to the same prefix with a reachable next hop. Which route does the router install, and what decides that?
MCQ-21A router learns 10.0.0.0/8 from OSPF with administrative distance 110 and from eBGP with administrative distance 20. Both next hops are reachable and no policy filters either route. Which route does the router install?
MCQ-8A router runs BGP and OSPF. BGP's best path for 10.20.0.0/16 has a next hop of 192.0.2.1, reachable via an OSPF route for 192.0.2.0/24 out of interface eth0. The BGP route is installed as the active route for the prefix. A packet arrives destined for 10.20.0.55. Which lookup actually decides the outgoing interface and rewritten next-hop MAC for this packet?
MCQ-22R1 has a route to 10.0.0.0/24 via 192.0.2.1 learned from OSPF. The OSPF neighbor is FULL and the route is in the RIB. Host A (10.0.0.10) sends 100 ICMP echo requests to Host B (10.0.0.20). R1's Gi0/0 output counter increments by 100, but a packet capture on Gi0/0 shows ARP requests for 192.0.2.1 and no ICMP echo packets. Which conclusion fits these observations?
MCQ-14During a design review, an engineer is asked to justify why BGP is classified as a path-vector protocol rather than a distance-vector or link-state one. Which property of BGP's advertisements is the reason?
MCQ-12Two routers on the same Ethernet segment show this on `show ip ospf neighbor`: the neighbor state cycles between EXSTART and EXCHANGE and never reaches FULL. `show ip ospf interface` on both shows area 0, hello 10 / dead 40, network type broadcast, and MD5 authentication with the same key. What is the cause?
MCQ-13A network runs OSPF with Area 0 as the backbone and Area 1 attached to it. Area 1 has two ABRs, R1 and R2, both with full adjacencies into Area 0. A new subnet is added inside Area 1 and advertised by R1 only. R2's Area 1 link fails, so R2 loses its only intra-area path to that subnet. R2 still has a healthy Area 0 adjacency. What does R2 do with traffic destined for the new subnet?
MCQ-11A router in OSPF area 0 has three neighbors on the same broadcast segment. Its neighbor table shows all three in Full state, but its link-state database contains only the router-LSAs it originated itself. What should you check first?
MCQ-10R1 summarizes the four prefixes 10.0.0.0/24 through 10.0.3.0/24 into a single 10.0.0.0/22 and advertises that /22 to R2. R1 has no default route and no discard route for the /22. The 10.0.2.0/24 link fails; R1 removes that /24 from its table but keeps advertising the /22 because the other three prefixes are still present. R2's only route to 10.0.2.0/24 is the /22 via R1. A host behind R2 runs `traceroute 10.0.2.7` and R1 is the last hop that replies. Which statement explains this?
MCQ-15AS 64500 runs iBGP in a full mesh and has eBGP sessions on router R1 to two upstreams, ISP-A (AS 64505) and ISP-B (AS 64520). Both advertise 203.0.113.0/24. R1's BGP table holds both routes, both next hops resolve, and the attributes are: | route | AS_PATH | MED | weight | local pref | | --- | --- | --- | --- | --- | | via ISP-A | 64505 64510 64515 | 0 | 0 | 100 | | via ISP-B | 64520 64530 | 0 | 0 | 100 | R1 installs the ISP-B route, and the other routers in AS 64500 do the same, so traffic to 203.0.113.0/24 leaves via ISP-B. The team wants every router in AS 64500 to send that traffic via ISP-A instead. Which change does that?
MCQ-16An edge router in AS 65001 peers with transit provider AS 65002. Its outbound route-map to AS 65002 is: ``` route-map TO-TRANSIT permit 10 match ip address prefix-list CUSTOMER-V4 route-map TO-TRANSIT deny 20 ``` CUSTOMER-V4 contains only 198.51.100.0/22. Customer AS 65010 originates 198.51.100.0/22 and, since last week, 203.0.113.0/24; both appear in the edge router's Adj-RIB-In from AS 65010. ROAs exist for both prefixes with origin AS 65010, and AS 65002's inbound policy permits any prefix whose origin AS is 65010. Looking glasses in AS 65002 and AS 65003 show 198.51.100.0/22 but no route for 203.0.113.0/24. What explains the missing /24?
MCQ-23R1 and R2 each run OSPF area 0 and EIGRP AS 10 and redistribute between the two protocols in both directions with no route tags or filters. Prefix 10.1.1.0/24 is connected to R3, which runs OSPF only and peers with R1. Before redistribution, R1 reached 10.1.1.0/24 through R3. After redistribution is enabled, traffic from R1 to 10.1.1.0/24 loops between R1 and R2 and never reaches R3. What causes the loop?
MCQ-9A leaf switch has four equal-cost 10 Gbps uplinks to a spine and hashes each flow on the 5-tuple. A backup job opens one TCP connection and sustains 9.3 Gbps while the other three uplinks carry about 0.2 Gbps each. The team wants the job to use more of the 40 Gbps without reordering packets. Which change achieves that?
MCQ-17A team runs two Linux routers, R1 and R2, on the same /24 LAN as the default gateway for 400 hosts. VRRP is configured with R1 as master (priority 200) and R2 as backup (priority 100), advert interval 1 s, and preempt enabled. R1's uplink to the core fails, so its interface toward the LAN stays up but it can no longer reach anything off-subnet. Hosts keep sending traffic to the virtual IP. What happens to the hosts' off-subnet traffic?
MCQ-18R1 and R2 are directly connected and run OSPF. A BFD session is up on the link, configured on both routers with a desired minimum transmit interval of 300 ms and a detection multiplier of 3; both sides negotiate the same values. OSPF's dead interval is 40 s. At t=0 the forwarding path between them fails while the interface stays up. What does BFD do?
MCQ-20A client behind a stateful firewall pair reaches a public HTTPS server. The pair runs active/active behind one NAT address: FW1 and FW2 each have their own uplink to the same ISP, ECMP hashes per flow, and the two firewalls do not synchronize session state. The client's SYN is hashed to FW1, which installs the NAT translation and forwards it; the server's SYN-ACK is hashed to FW2. The client retransmits the SYN until the connection times out. What is the most likely cause?
MCQ-19A VXLAN overlay runs over an underlay where every link has an interface MTU of 1500 bytes: the standard Ethernet limit, so the largest frame on the wire is 1514 bytes including the 14-byte Ethernet header. VTEPs set the DF bit on the outer IPv4 header, no jumbo frames are configured, and ICMP is filtered in the underlay so path MTU discovery cannot work. Overlay hosts keep their interface MTU at 1500 and send untagged frames. Bulk TCP transfers between two overlay hosts stall after a few hundred kilobytes, while small requests complete normally. What is the largest inner IP packet the overlay can carry end to end without the outer frame exceeding the underlay's 1500-byte interface MTU?
MCQ-24A 40-router OSPF area carries one service whose p99 latency rose from 12 ms to 45 ms after a maintenance window and stayed there for six hours. Interface error counters on every link are zero, link utilization peaks at 30%, and router CPU sits at 25%. Which measurement would confirm the latency is caused by control-plane churn rather than a data-plane fault?