JNCIA-Design flashcards

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

Collapsed core

A design that merges the core and distribution layers into one device pair, reducing cost and hop count when the site does not need a dedicated high-speed core.

Availability formula

Availability equals MTBF divided by MTBF plus MTTR. A 99.99% target permits roughly 52 minutes of unplanned downtime across a full year.

Common-mode failure

A single fault that disables multiple supposedly redundant components, such as one power feed, one conduit, or one upstream switch shared by both paths.

Route summarization benefit

Advertising an aggregate prefix instead of many specific prefixes shrinks routing tables, speeds convergence, and hides topology churn inside a site from the rest of the network.

Jumbo frame MTU

An Ethernet frame size commonly set to 9000 bytes for storage and data center traffic. Every L2 and L3 hop must support the same MTU or packets are dropped.

Three-tier hierarchy

Access attaches endpoints, distribution aggregates and applies policy, and core forwards at high speed. Each layer has a distinct role and a distinct scaling concern.

RSTP versus STP

RSTP (802.1w) converges in sub-second to a few seconds using edge ports and proposal-agreement, while legacy 802.1D STP can take 30 to 50 seconds.

PoE budget calculation

Sum the per-port power class draw across all powered devices, then compare with the switch power supply capacity, leaving headroom for future ports and cable loss.

STP root placement

Configure the root bridge on the distribution or core pair using root primary and secondary commands, keeping the root stable and near the center of the topology.

BPDU guard

A feature that error-disables an edge port if it receives a BPDU, preventing an unauthorized switch from joining the topology and disrupting the spanning tree.

Link aggregation hashing

LACP bundles distribute traffic per flow using a hash of addresses and ports. One conversation uses one member link, so single-flow throughput never exceeds one link.

First-hop redundancy protocols

VRRP is standards-based, HSRP is Cisco-proprietary, and GLBP adds load sharing. All present a virtual IP and MAC shared by a router pair.

Wireless channel planning

In 2.4 GHz use only channels 1, 6, and 11 in North America. Plan 5 GHz cells for about -67 dBm RSSI at the edge and 20-25 dB SNR for voice.

Voice VLAN design

A separate VLAN for IP phones, assigned via LLDP-MED or CDP, with QoS marking trusted at the access edge and priority queuing for signaling and RTP.

VLAN-to-subnet mapping

Each VLAN maps to exactly one IP subnet. Inter-VLAN routing happens at distribution or core, and VLANs are confined to a single distribution block.

BGP best-path order

Weight, then LOCAL_PREF, then locally originated, then shortest AS_PATH, then origin type, then MED, then eBGP over iBGP, then lowest IGP metric, then lowest router ID.

BGP max-prefix limit

A per-neighbor cap on accepted prefixes that shuts or restricts the session when exceeded, protecting against route leaks and misconfiguration from a peer.

Multi-homing resilience

Only multi-homing across two distinct providers survives a provider outage. Dual links to one provider protect against link failure but not provider failure.

Tunnel MSS clamping

Reducing TCP maximum segment size on tunnel interfaces so encapsulated packets fit the path MTU, avoiding black holes when ICMP fragmentation-needed messages are filtered.

IPsec overhead

ESP in tunnel mode adds roughly 50 to 73 bytes per packet depending on algorithms. GRE adds 24 bytes. Both reduce usable payload on a 1500-byte path.

SD-WAN overlay design

Encrypted tunnels over Internet and MPLS transport, application-aware path selection, centralized policy, and local Internet breakout for SaaS and cloud traffic.

QoS trust boundary

The point where markings are accepted or reset, typically the access switch for phones and the WAN edge for untrusted sources. Markings beyond it are trusted.

Spine-leaf topology

Every leaf connects to every spine and leaves never connect to each other. All paths are equal cost, enabling ECMP active-active forwarding and predictable hop count.

Oversubscription ratio

Host-facing bandwidth divided by uplink bandwidth. A 3:1 ratio means hosts can demand three times the uplink capacity; 1:1 is non-blocking.

Bisection bandwidth

The total capacity crossing an imaginary line splitting the fabric in half. It bounds worst-case traffic between halves and equals spine links times link speed.

VXLAN VNI space

VXLAN uses a 24-bit VNI supporting about 16 million segments, compared with the 12-bit VLAN ID limit of 4094 usable segments.

EVPN route types

Type 2 advertises MAC and IP bindings, type 3 handles inclusive multicast for BUM traffic, and type 5 advertises IP prefixes for routing between subnets.

Lossless Ethernet for storage

FCoE and RoCE require priority flow control and enhanced transmission selection so congestion drops do not occur, since dropped storage frames cause severe latency and retransmission.

vPC and MLAG

Two switches appear as one logical device to attached hosts and downstream switches, allowing both uplinks to forward without spanning-tree blocking.

DMZ placement

Public-facing servers sit in a screened subnet between two firewall interfaces. Inbound sessions terminate there, and internal access is brokered rather than direct.

802.1X roles

The supplicant is the endpoint, the authenticator is the switch port, and the authentication server is RADIUS. EAP-TLS uses certificates on both sides.

IPsec ESP versus AH

ESP encrypts and authenticates payloads and survives NAT traversal. AH authenticates the whole packet including the IP header but cannot pass through NAT.