Introduction

Within this post, we will show you how to build a virtualized NSX SDN lab.

Our lab will be loosely based around the VMware NFV 2.0.1 Reference Architecture.
For those of you who haven't worked with the VMware NFV reference architecture, it consists of a set of validated VMware software components and high-level design reference.

This guide assumes you have a good understanding of NSX. However, if this isn't the case, then I highly recommend the VMware NSX Technical Deep Dive from IPSpace.net
The webinar describes VMware NSX principles, architecture, and components, including overlay virtual networking, physical-to-virtual gateways, network services, and security. CLICK HERE to join.

Note: With regards to NSX licensing, checkout VMUG for personal/non-production licensing: https://www.vmug.com/Join/EVALExperience

Summary of Steps

The lab will be built based on the following steps:

  1. Building the VMware Infra (ESXi hosts and vDS's).
  2. Deploy the NSX Manager via OVF.
  3. Register the NSX Manager with the vCenter server.
  4. Deploy the NSX Control Cluster.
  5. Deploy VXLAN (via VIB) to the hosts.
  6. Verify connectivity between 2 VMs across a logical switch.

Lab Details

Overview

The lab will consist of 2 x physical ESXi servers. The first ESXi server will host the management components, and the other ESXi server will host a number of virtualized ESXi servers used for compute and Edge (ESG's). ESGs being VMs that will perform breakout from the overlay (VXLAN) to the underlay.

image2-11
Figure 1 - Lab overview.

Components

Our lab will be based on the following components.

Physical Host VM Nested VMs vCPU Memory HDD
Management ESXi vCenter 2 4Gb 120GB
NSX Manager 4 16Gb 60GB
Compute ESXi ESXi001 2 8Gb 40GB
ESXi002 2 8Gb 40GB
ESXi003 2 8Gb 40GB
ESXi004 2 8Gb 40GB
ESXi005 2 16Gb 100GB
ESXi006 2 16Gb 100GB
NSX Controller (x3) 4 4Gb 28GB

vSphere Setup

First of all, we will build the vSphere infra. This involves installing the ESXi hosts, creating the required clusters and building our vDS's.

Requirements

Before we begin the following requirements much be performed.

  • ESXi installed on each of the physical hosts.
  • vCenter installed.

Furthermore, the following services must be available:

  • DNS for internal resolution of VMware components i.e VC etc.
  • NTP for time synchronization.

Create Clusters

From within vCenter create the following clusters,

  • NFV-Data
  • NFV-Control
  • NFV-Edge

Build ESXi Hosts

Create 6 virtual ESXi VMs.

  • 8Gb RAM, 2 x vCPUs and 100GB disk.
  • Enable VT-x under the CPU section.
  • Create 2 x NICs on each VM.
  • Assign a trunk based port group to each interface.
  • Upon the vSS on the physical host used for compute, set promiscuous mode to enabled.

image8-1
Figure 2 - Create ESXi VM's.

Once you've created them, install ESXi6.5 on each. Once installed:

  • Configure DNS resolution.
  • Configure the following IP and VLAN tags against the management interface, based on the below.
ESX VM IP/MASK VLAN Cluster
ESXi001 172.29.111.1 111 NFV-Control
ESXi002 172.29.111.2 111 NFV-Control
ESXi003 172.29.111.3 111 NFV-Data
ESXi004 172.29.111.4 111 NFV-Data
ESXi005 172.29.121.5 121 NFV-Edge
ESXi006 172.29.121.6 121 NFV-Edge

Update Clusters

Next, import each of the hosts into vCenter. Resulting in the following vSphere topology:
image11

vSphere Networking

Now, at this point, each ESXi host VM will be only using a single VMNIC and the virtual standard switch (vSS). Therefore our next action will be to migrate the ESXi over to a vDS and to use both VMNIC uplinks. To do this, within the network section of vCenter go to Add and Manage Hosts. This will start a wizard where you can attach your hosts, assign uplinks and configure the required VLANs (shown below).

vDS PortGroup VLAN
vDS-Resource-Infra PG-Resource-Infra-Mgmt 111
PG-Resource-Infra-VMMgmt 112
vDS-Edge-Infra PG-Edge-Infra-Mgmt 121
PG-Edge-Infra-VMMgmt 122

image7
Figure 3 - vSphere networking overview.

What's Next?

Now that we have the vSphere infra up and running, we can look to start the NSX installation, covered within part 2 of this 4 part series.