CCNP Service Provider flashcards

32 cards. Read the term and the definition here. The deck at the top of this page is optional.

Access-aggregation-core hierarchy

Three-tier SP topology: access terminates subscriber lines, aggregation concentrates traffic and applies BNG/QoS policy, and core transports label-switched traffic between PoPs at the highest capacity.

PoP redundancy rule

A point of presence should keep subscriber sessions alive after losing one core router and one inter-PoP link, so each core router needs at least two diverse uplinks to different backbone paths.

IS-IS in SP cores

IS-IS runs directly on Layer 2 without IP, scales through L1/L2 hierarchy, and extends via TLVs, which is why large service provider cores favor it over OSPF for IGP transport.

BGP route reflector purpose

A route reflector removes the need for full iBGP mesh by re-advertising routes between clients and non-clients, using CLUSTER_LIST and ORIGINATOR_ID to prevent loops and preserve origin.

QoS forwarding class design

Service provider designs typically use five to eight forwarding classes, reserving strict priority for real-time traffic only and giving best-effort and scavenger classes lower weights.

NSR versus NSF

Nonstop routing keeps the control-plane session state on a standby supervisor so failover does not reset BGP or IGP peers, while nonstop forwarding relies on peer graceful-restart helpers to keep forwarding alive.

BGP best-path weight

Weight is the first comparison in Cisco best-path selection, is locally significant, and is never carried in BGP updates, so it can override every other attribute including local preference.

BGP local preference scope

Local preference is exchanged only among iBGP peers and is the second tiebreaker, so it steers outbound path choice for an entire autonomous system rather than a single router.

MED comparison default

Multi-exit discriminator is compared only among paths received from the same neighboring AS, and a missing MED is treated as zero unless missing-as-worst is configured.

iBGP split horizon

A route learned from an iBGP peer is not advertised to another iBGP peer, so internal BGP requires either a full mesh or route reflectors to propagate routes across the AS.

Route reflector client rule

A route received from a client is reflected to all peers except the originating client, while a route from a non-client goes to clients and non-clients, with next hop unchanged.

OSPF LSA type 5

Type 5 external LSAs are originated by ASBRs and flooded throughout normal areas to describe redistributed routes, but they are not allowed inside stub or NSSA areas.

IS-IS ATT bit

An L1/L2 router sets the attached bit in its L1 LSP so L1 routers install a default route toward the nearest L2 exit point instead of learning external prefixes.

Penultimate hop popping

The penultimate router pops the transport label and forwards the packet unlabeled, which is why the egress PE advertises implicit-null for its loopback prefix to LDP.

LDP-IGP synchronization

The IGP adjacency on a link is held down until LDP converges on that same link, preventing a window where labeled traffic is dropped because the label binding is missing.

Loop-free alternate condition

A neighbor N is an LFA for destination D if its distance to D is less than the local distance to D plus the local distance to N, which guarantees the alternate path cannot loop back.

MPLS LFIB lookup

The label forwarding information base maps an incoming label and interface to an outgoing label and interface, and the incoming label is used as the lookup key instead of the IP destination.

Route distinguisher function

An eight-byte RD prepended to an IPv4 prefix makes it unique across the MPLS domain, allowing overlapping customer addresses to coexist in separate VRFs on the same PE.

Route target import export

Route targets are extended communities attached on export and matched on import, so a VRF only installs VPN routes whose RT matches one of its configured import targets.

L3VPN label stack order

The ingress PE pushes the VPN label first and the transport label on top, so P routers swap only the outer transport label and the egress PE forwards on the inner VPN label.

VPNv4 address family

MP-BGP address family 1 with subsequent address family 128 carries VPNv4 routes, including the RD, route targets, MPLS label, and a next hop rewritten to the PE loopback.

Inter-AS Option B

ASBRs run MP-eBGP carrying VPNv4 routes with next-hop-self, so each ASBR holds all VPN routes and forwards labeled packets, unlike Option A which uses per-VRF interfaces.

Inter-AS Option C

P routers exchange multihop MP-eBGP VPNv4 directly between ASes while ASBRs only redistribute labeled IPv4 routes, keeping VPN route state off the ASBRs.

VPLS forwarding behavior

VPLS emulates a LAN by building a full mesh of pseudowires between PEs and learning MAC addresses in the data plane, flooding unknown unicast, broadcast, and multicast across the mesh.

EVPN route type 2

Type 2 MAC/IP advertisement routes distribute customer MAC addresses and optionally IP addresses, enabling control-plane MAC learning, ARP suppression, and multihomed active-active forwarding.

EVPN designated forwarder election

On a multihomed Ethernet segment, PEs elect one designated forwarder per segment to send broadcast and unknown unicast traffic, preventing duplicate delivery to the customer site.

RSVP-TE CSPF

Constrained shortest path first uses IGP traffic engineering extensions such as available bandwidth and admin groups to compute a path that satisfies a tunnel's constraints before RSVP signals it.

NETCONF transport and port

NETCONF runs over SSH on TCP port 830 and uses XML-encoded RPCs such as get-config, edit-config, lock, commit, and discard-changes against a candidate datastore.

RESTCONF HTTP verbs

RESTCONF maps YANG data to HTTP: GET reads, POST creates, PUT replaces, PATCH merges, and DELETE removes, with JSON or XML payloads over HTTPS on port 443.

gNMI operations

gNMI carries protobuf messages over HTTP/2 and supports Capabilities, Get, Set, and Subscribe, where Subscribe streams telemetry on-change, at a sample interval, or on poll.

Model-driven telemetry

The device pushes YANG-modeled operational data to a collector at a configured interval or on change, replacing periodic SNMP polling with higher-resolution, structured, and scalable streams.

IP SLA tracking object

An IP SLA probe measures latency, jitter, or reachability, and its tracking object can withdraw a static route or shift traffic when a configured threshold is breached.