Deployment Guide

11 ScaleIO/VxFlex OS IP Fabric Best Practice and Deployment Guide with OS10EE | version 1.0
The following physical concepts apply to all routed Leaf-Spine topologies:
• Each leaf switch connects to every spine switch in the topology.
• Spine switches only connect to leaf switches.
• Leaf switches connect to spine switches and other devices such as servers, storage arrays, and edge
routers.
• Servers, storage arrays, edge routers and other non-leaf-switch devices never connect to spine
switches.
• It is a best practice to use VLT for connecting leaf switch pairs. This provides redundancy at the layer
2 level for devices that use an active-active link aggregation group (LAG) to attach to both switches.
2.1 Management network
A management network is not a requirement to setup or configure the Leaf-Spine network. However, Dell
EMC recommends using a management network in larger network topologies for efficient management of
several devices. Typical deployments use a single management-traffic network isolated from the production
network. A Dell EMC S3048-ON switch in each rack can provide connectivity to the management network.
Out-of-band (OOB) ports on each production switch connect them to the management network.
For a detailed configuration example see ScaleIO IP Fabric Best Practice and Deployment Guide
.
Management network
2.2 IP-based storage
One of the benefits of using VxFlex OS in a Leaf-Spine infrastructure is its reachability. Any rack in the data
center can reach the solution using IP routing. This allows VxFlex OS to reach hundreds of SDS-SDC nodes
with nearly seamless scale-out for growth.