I85PREP
⌘ K
CCNP Service Provider study materials
CCNP Service Provider study resources

CCNP Service Provider: Service Provider Network Architecture

CCNP Service Provider · Work through your exam, one topic at a time.

All guide sections

Service Provider Network Architecture

Domain 1 of SPCOR (15%) examines how a service provider network is assembled: functional planes, PoP and tier structure, hardware roles, network virtualization, and high-availability mechanisms. The questions are conceptual, but they turn on precise distinctions, so define each term before comparing options.

Functional planes

  • Control plane: routing protocols (IS-IS, OSPF, BGP), LDP, RSVP-TE; runs on the route processor (RP) or supervisor and builds the RIB.
  • Data (forwarding) plane: line cards and switch fabric; uses the FIB and LFIB. On modular chassis, forwarding is distributed to line-card CPUs/ASICs.
  • Management plane: CLI, SNMP, NETCONF/RESTCONF, syslog, and telemetry; place it out-of-band or protect it with ACLs.
  • CoPP polices traffic punted to the RP. It protects the control plane, not customer traffic in the data plane.

PoP and tier structure

  • Access: customer handoff and subscriber aggregation; aggregation: concentrates access nodes and is often the L2/L3 boundary.
  • Edge/PE: VRFs, MPLS, QoS, and peering; core/P: label swapping only.
  • Route reflectors are control-plane devices and do not have to sit in the forwarding path. A dedicated RR can be a P router or a separate appliance.
  • POP redundancy uses dual edge routers, dual uplinks, and diverse fiber.

Hardware and virtualization

  • Modular chassis contain RPs, line cards, fabric modules, and power; ASR 9000 and NCS 5500 families run Cisco IOS XR.
  • Network virtualization (segment routing, EVPN, VXLAN, NFV) improves multi-tenancy and service agility and reduces protocol sprawl.
  • NFV virtualizes network functions; SDN centralizes control. They are related but not the same.

High availability

  • NSF/graceful restart relies on helper routers; NSR keeps TCP sessions across an RP switchover.
  • BFD, TI-LFA, MPLS FRR, RP redundancy, and fabric redundancy address different failure domains.

Worked example

A PE's customer-facing interface is hit by a flood of BGP updates and ARP requests during a loop. Which plane is at risk? Packets destined to the RP are punted, so without CoPP the RP CPU saturates and BGP/IGP sessions drop even though line cards can still forward existing flows. The fix is a CoPP policy with classified policers, plus uRPF and BGP max-prefix. The data plane may survive briefly, but new flows fail once the control plane reconverges.

Traps

  • Treating SNMP as control plane; it is management.
  • Confusing graceful restart with NSR.
  • Assuming an RR must forward traffic.
  • Putting VRFs on P routers that only swap labels.
  • Assuming every chassis forwards centrally.