2024-09-10

Network Architecture

networkspine-leafdata centerCCNAarchitecture

>Redundancy is important while building a network.
>Two is one, one is none.

Two tier / collapsed core

>The core switch is collapsed into the distribution switch.
>Used by businesses of smaller scale, or businesses on a budget.
>It's not about how many employees we have, it's about how many buildings a business has.
>Does all kinds of routing, packet switching, QoS.

A two-tier network where the core and distribution layers are collapsed into a single switch tier

Three tier

>The network spans multiple buildings.
>The core does nothing except transfer data as fast as possible.
>In the case of just a couple of buildings we can use LAG / EtherChannels, but as the number of buildings grows it makes the network messier, so we use three-tier architecture.

A three-tier network with separate core, distribution and access layers spanning multiple buildings

Spine-leaf architecture (east-west traffic)

>Mostly used in data centers.
>The spine is the backbone; servers connect to the leaves.
>Full mesh between spine and leaf.
>Very fast traffic flow — combines traffic.

A spine-leaf topology where every leaf switch connects up to every spine switch

LAG (link aggregation) interface / EtherChannel — taking bandwidth from multiple links and bundling them together.

ASIC — Application-Specific Integrated Circuit.

Architecture recall1/3

In a two-tier (collapsed core) design, what happens to the core layer?