Manualshelf

Deployment Guide

ManualsBrandsDell ManualsConverged InfrastructureEnterprise Solution Resources
12345678910
10 Leaf-Spine Deployment and Best Practices Guide | Version 1.0
3 Leaf-spine overview
The connections between leaf and spine switches can be layer 2 (switched) or layer 3 (routed). The terms
“layer 3 topology” and “layer 2 topology” in this guide refer to these connections. In both topologies,
downstream connections to servers, storage and other endpoint devices within the racks are layer 2 and
connections to external networks are layer 3.
The following concepts apply to layer 2 and layer 3 leaf-spine topologies:
• Each leaf switch connects to every spine switch in the topology.
• Servers, storage arrays, edge routers and similar devices always connect to leaf switches, never to
spines.
The layer 2 and layer 3 topologies each use two leaf switches at the top of each rack configured as a Virtual
Link Trunking (VLT) pair. VLT allows all connections to be active while also providing fault tolerance. As
administrators add racks to the data center, two leaf switches configured for VLT are added to each new rack.
The total number of leaf-spine connections is equal to the number of leaf switches multiplied by the number of
spine switches. Bandwidth of the fabric may be increased by adding connections between leafs and spines
as long as the spine layer has capacity for the additional connections.
3.1 Layer 3 leaf-spine topology
In a layer 3 leaf-spine network, traffic between leafs and spines is routed. The layer 3 / layer 2 boundary is at
the leaf switches. Spine switches are never connected to each other in a layer 3 topology. Equal cost multi-
path routing (ECMP) is used to load balance traffic across the layer 3 network. Connections within racks from
hosts to leaf switches are layer 2. Connections to external networks are made from a pair of edge or border
leafs as shown in Figure 7.
Layer 3 leaf-spine network