Deployment Guide

9 ScaleIO/VxFlex OS IP Fabric Best Practice and Deployment Guide with OS10EE | version 1.0
Meta Data Manager (MDM)
The MDM configures and monitors VxFlex OS. It contains all the metadata required for VxFlex OS operation.
Configure the MDM in Single Mode on a single server or redundant Cluster Mode – three members on three
servers or five members on five servers.
Note: Use Cluster Mode for all production environments. Dell EMC does not recommend Single Mode
because it exposes the system to a single point of failure.
1.2 Traffic model for the modern data center
The legacy traffic pattern for data centers has been the classic client-server path and model. The end user
sends a request to a resource inside the data center and that resource computes and responds to the end-
user traffic. The data center is designed more for the expected North-to-South traffic that travels to and from
the data center, rather than the possible East-to-West traffic that traverses between the racks in the data
center. The design focuses on transport into and out of the data center and not residential applications. This
results in traffic patterns that do not always follow a predictable path due to asymmetrical bandwidth between
the networking layers.
The evolution that has taken place is the increased machine-to-machine traffic inside the data center. The
reasons for this increased East-West traffic are:
• Applications are much more tiered where web, database and storage interact with each other.
• Increased virtualization where applications easily move to available compute resources.
• Increased server-to-storage traffic due to storage solutions like Dell EMC VxFlex OS that involve
hundreds of deployed nodes and require higher bandwidth and scalability.
In large data clusters, VxFlex OS can achieve distribution and replication of all the data. This places a heavy
burden on the data center infrastructure since all the data is replicated and that replication happens
continuously. This is a clear difference from the old model of one interaction being limited to one North-South
communication. Instead, there is a noticeable increase of East-West traffic before any messages are
delivered to the end user.
A well thought out design includes predictable traffic paths and well-used resources. By industry consensus, a
Leaf-Spine design with an ECMP, load-balanced mesh creates a well-built fabric that both scales and
provides efficient, predictable traffic flows. A two-tier Leaf-Spine design provides paths that are never more
than two hops away, which results in consistent round-trip time.
1.3 Attachments
This main document includes multiple attachment files for spine-switch configurations and leaf-switch
configurations. The example configurations are based on a solution assembled in the Dell EMC Networking
lab and require editing to fit any other infrastructure.