CCIE CCIE Network Design 2 — Questions and Answers
Question 1: When designing a BGP route reflector (RR) cluster, which attribute prevents routing loops between route reflectors in the same cluster?
- ORIGINATOR_ID
- CLUSTER_LIST (Correct answer)
- LOCAL_PREF
- COMMUNITY
Correct answer: CLUSTER_LIST
CLUSTER_LIST is appended by each RR as it reflects routes; a router that sees its own cluster ID in the CLUSTER_LIST discards the update, preventing intra-cluster loops.
Question 2: A network designer is planning IP address space for a large enterprise with 500 subnets. Which practice best supports route summarization to minimize routing table size at the core?
- Assign subnets randomly from the available pool
- Allocate contiguous IP blocks aligned to summarization boundaries per site (Correct answer)
- Use /30 point-to-point links everywhere
- Deploy NAT at every distribution layer router
Correct answer: Allocate contiguous IP blocks aligned to summarization boundaries per site
Allocating contiguous, boundary-aligned IP blocks per region or site allows a single summary route to represent hundreds of subnets, minimizing core routing table entries.
Question 3: In an MPLS L3VPN design, which component is responsible for maintaining per-VPN routing tables and distributing VPN routes using MP-BGP with VPNv4 address families?
- P router (Provider core)
- PE router (Provider Edge) (Correct answer)
- CE router (Customer Edge)
- RR for the underlay IGP
Correct answer: PE router (Provider Edge)
PE routers maintain VRF tables per customer VPN, import/export routes using Route Distinguishers and Route Targets, and exchange VPNv4 prefixes with other PEs via MP-BGP.
Question 4: A designer must ensure that a campus network can recover from a distribution-layer switch failure in under 1 second. Which combination of technologies best achieves this?
- HSRP with default timers and Rapid PVST+
- StackWise Virtual with ECMP and BFD-triggered BGP (Correct answer)
- VRRP with 3-second hello timers
- STP PortFast on uplinks with HSRP preempt
Correct answer: StackWise Virtual with ECMP and BFD-triggered BGP
StackWise Virtual eliminates the virtual gateway failover step entirely, while BFD provides sub-second failure detection and ECMP uses all uplinks simultaneously for instant path switching.
Question 5: Which QoS queuing mechanism is recommended for voice traffic in an enterprise WAN design to guarantee bandwidth and minimize jitter?
- FIFO queuing
- Weighted Fair Queuing (WFQ)
- Low Latency Queuing (LLQ) (Correct answer)
- Custom Queuing (CQ)
Correct answer: Low Latency Queuing (LLQ)
LLQ combines Priority Queuing for strict latency-sensitive traffic (like voice) with class-based WFQ for other classes, guaranteeing bandwidth while preventing queue starvation.
Question 6: When designing OSPF for a large enterprise with 500+ routers, which design choice most effectively reduces CPU and memory consumption on all routers?
- Use OSPF area 0 for all routers
- Segment routers into multiple areas with summarization at ABRs (Correct answer)
- Enable OSPF fast hello packets globally
- Configure all non-DR routers as stub routers
Correct answer: Segment routers into multiple areas with summarization at ABRs
Multiple areas restrict LSA flooding to within each area, and ABR summarization reduces the size of the routing table and LSDB on all non-ABR routers, cutting CPU and memory usage.
Question 7: A company requires that customer branch offices connect to headquarters using encrypted tunnels over the internet with hub-and-spoke topology but dynamic spoke-to-spoke tunnels on demand. Which Cisco technology fits this design?
- DMVPN (Dynamic Multipoint VPN) (Correct answer)
- Static point-to-point GRE tunnels
- IPsec tunnel mode only
- MPLS L2VPN with pseudowires
Correct answer: DMVPN (Dynamic Multipoint VPN)
DMVPN uses NHRP to allow spoke routers to dynamically discover each other's tunnel endpoints and build direct IPsec tunnels without pre-configuration, supporting dynamic spoke-to-spoke communication.
CCIE CCIE Network Design 2
When designing a BGP route reflector (RR) cluster, which attribute prevents routing loops between route reflectors in the same cluster?