Network Engineer interview questions

Static routing offers explicit control and low protocol overhead, while dynamic routing adapts reachability as topology changes.

• Static routes suit stable paths, defaults, and tightly controlled exceptions but require manual lifecycle management.
• Dynamic protocols discover alternate paths and converge after failures at the cost of control-plane traffic and complexity.
• Many networks combine both, with redistribution or preference rules defining how the route sources interact.

The RIB contains candidate and selected routes from routing sources, while the FIB contains optimized entries used to forward packets.

• Routing protocols and static configuration contribute routes to the control-plane RIB.
• The best routes and resolved next hops are programmed into the data-plane FIB.
• A platform may keep several candidate paths in the RIB even though only selected paths reach the FIB.

Forwarding uses the longest matching prefix, while route installation for an identical prefix compares source preference and then protocol metrics.

• A more specific /24 is used instead of a covering /16 regardless of the /16 source's lower administrative distance.
• For the same prefix, administrative distance selects between sources such as static, OSPF, and EIGRP.
• The chosen protocol's metric ranks paths learned through that protocol, with equal-cost paths potentially installed together.

OSPF floods link-state advertisements within an area so routers can build a synchronized topology database and run SPF.

• LSAs describe routers, links, networks, and reachability rather than directly sending a complete route table as distance-vector updates.
• Each router independently runs Dijkstra's algorithm from itself as the root.
• The resulting shortest-path tree supplies next hops and costs for routes installed into the routing table.

OSPF neighbors need compatible interface and area parameters before their state machines can reach Full where adjacency is expected.

• Area ID, authentication, hello and dead timers, and key network parameters must be compatible.
• Router IDs must be unique, and IP connectivity must exist on the shared link.
• Stub-area flags and MTU behavior can also prevent successful database exchange even after Hellos are seen.

OSPF elects a designated router and backup to reduce adjacency count and LSA exchange on multiaccess segments.

• Other routers form full adjacencies with the DR and BDR instead of every router fully adjoining every other router.
• Election uses interface priority first and router ID as a tie-breaker.
• The election is non-preemptive, so a later router with a higher priority does not automatically replace the current DR.

OSPF areas limit detailed topology flooding and SPF scope, while area border routers exchange reachability between areas.

• Routers inside an area share that area's LSDB rather than every link-state detail in the entire domain.
• ABRs connect non-backbone areas to area 0 and maintain separate area databases.
• Summarization at area boundaries can reduce route count and isolate some topology changes.

OSPF adds outgoing interface costs along a path and prefers the route with the lowest total cost.

• Cost is commonly derived from reference bandwidth divided by interface bandwidth.
• The reference bandwidth must be set consistently so modern high-speed links remain distinguishable.
• Administrators can set interface cost explicitly when policy should differ from the bandwidth-derived value.

EIGRP uses DUAL to select a successor as the best loop-free path and may retain feasible successors as safe backups.

• The successor has the lowest feasible distance for the destination.
• A neighbor qualifies as a feasible successor when its reported distance is lower than the current feasible distance.
• A feasible successor can be installed quickly after failure without first running a network-wide query.

eBGP exchanges routes between autonomous systems, while iBGP distributes BGP routes within one autonomous system.

• eBGP adds the local AS to AS_PATH when advertising routes across an AS boundary.
• iBGP does not advertise routes learned from one iBGP peer to another by default, requiring full mesh or route reflectors.
• Their default session assumptions and next-hop handling differ, although policy can modify much of the behavior.

BGP evaluates ordered attributes and policy, commonly including weight, local preference, locally originated status, AS_PATH, origin, MED, and eBGP versus iBGP.

• Higher local preference is favored throughout an AS, while shorter AS_PATH is generally preferred later in the process.
• Lower MED can suggest a preferred entry point to a neighboring AS under comparable conditions.
• The exact order and vendor-specific attributes must be verified for the platform in use.

Local preference tells routers inside an AS which exit to prefer, while MED suggests to a neighboring AS which entry point to use.

• Higher local preference wins and is propagated to iBGP peers within the AS.
• Lower MED is preferred and is normally compared among paths received from the same neighboring AS.
• Local preference directly controls outbound choice, whereas MED is a suggestion the neighboring network may ignore or override.

Redistribution can create loops, feedback, suboptimal paths, and inconsistent reachability when different protocol semantics meet.

• Metrics must be translated because OSPF, EIGRP, and other protocols measure paths differently.
• Route tags and filtering help prevent routes from being redistributed back into their source.
• Redistribution points should be few, deliberate, documented, and paired with explicit prefix policy.

MPLS separates transport labels from service context so a provider can carry isolated customer routes across a shared core.

• Provider-edge routers assign customer routes to VRFs and attach VPN labels.
• Core label-switching routers forward transport labels without needing customer route tables.
• MP-BGP distributes VPN routes and labels between provider edges while route targets control import and export.

STP elects the switch with the lowest bridge ID as root, then each non-root switch selects its lowest-cost path toward that root.

• Bridge ID combines bridge priority, extended system ID, and MAC address.
• Root path cost is the sum of port costs along the path.
• Ties are resolved through ordered BPDU fields such as sender bridge ID and sender port ID.

STP assigns root and designated roles while keeping redundant non-designated or alternate ports non-forwarding.

• Each non-root switch has one root port providing its best path toward the root bridge.
• Each segment has one designated port providing the best path from that segment toward the root.
• Redundant ports that would create a loop remain non-forwarding and can become active after a topology change.

RSTP uses explicit proposal and agreement handshakes and simplified port states to activate safe links more quickly.

• It uses discarding, learning, and forwarding states instead of the longer classic sequence.
• Alternate ports provide immediately known backup paths toward the root.
• Edge ports can enter forwarding directly when they are known not to connect another switch.

BPDU Guard protects edge ports from unexpected switches, while Root Guard prevents a port from accepting a superior root claim.

• BPDU Guard commonly disables an edge port that receives a BPDU.
• Root Guard places the port into a root-inconsistent state while superior BPDUs continue.
• Both features enforce topology intent but apply to different trust assumptions and port roles.

Allowed VLANs define which tagged VLANs cross a trunk, while the native VLAN defines handling for untagged frames.

• Pruning unused VLANs limits unnecessary broadcast reach and accidental exposure.
• Native VLAN mismatches can place untagged traffic into different broadcast domains at each end.
• Both sides should use an explicit, consistent native VLAN policy and matching allowed lists.

EtherChannel members must have compatible physical and Layer 2 settings so they operate as one logical link.

• Speed, duplex, access or trunk mode, native VLAN, and allowed VLANs must agree.
• The channel protocol and group configuration must match the intended bundle.
• STP sees the port-channel as one logical link, preventing individual members from being blocked separately.