VCP6-NV Study Notes – Section 2 Understand VMware NSX Physical Infrastructure Requirements – Part 2

Objective 2.2: Determine Physical Infrastructure Requirements for a VMware NSX Implementation

Discern management and edge cluster requirements



  • Management Cluster
    • Composed by including vCenter Server, NSX manager, NSX controller, Cloud Management Systems (CMS), and other shared components.
    • Compute and memory requirements are pre-identified by the sum of the  minimum supported configuration
    • Enabling LACP is possible to improve resiliency  of management components.
    • Don’t require VXLAN provisioning (however in small design where Edge and Management cluster are collapsed, still become necessary)
  • Compute Cluster
    • Are designed with the following considerations
      • Host density per rack and automation dependences
      • Availability and Mobility of Workload
      • Connectivity –> single VTEP vs multi VTEP
      • Topology implication and IP processing for VTEP and vmkernel
    • Lifecycle and workload drivers consideration
      • Growth and changes
      • Multi-rack, zoning
      • Same 4 VLANs are required for each rack (Management, vMotion, Storage, VXLAN)
    • Workload centric allocation, compliance SLA be met via:
      • Cluster separation
      • Separate VXLAN and transport zone
      • Per tenant DLR and Edge routing domains
      • DRS and resources reservation
  • Edge cluster
    • provide on-ramp and off-ramp connectivity to physical world
    • Allow communication with physical devices connected in NSX L2 bridge
    • Host and control DLR routing
    • May have centralized logical or physical services
    • NSX Controllers can be hosted in an edge cluster –> dedicated vCenter is used to manage compute and edge resources
    • Edge cluster resources have anti-affinity requirement to protect the active-standby configuration or maintain bandwidth availability during failure
    • Needs and Characteristics
      • Edge VM is CPU centric with consistent memory requirements
      • Additional VLAN are required for North-South routing and bridging
      • Recommended teaming option for Edge hosts is “route based on SRC-ID”. Use of LACP is highly discouraged due to vendor specific requirements of route peering over LACP


Differentiate minimum/optimal physical infrastructure requirements for a VMware NSX implementation

  • System requirements for NSX: You can install one NSX Manager per vCenter Server, one instance of Guest Introspection and Data Security per ESXi™ host, and multiple NSX Edge instances per datacenter
    • Hardware (or virtual hardware)
      • NSX Manager –> 16 GB (24 for certain deployment sizes)RAM , 4 vCPU (8 in certain sizes), 60 GB vmdk space
      • NSX Controller –> 4 GB RAM, 4vCPU, 20 GB vmdk space
      • NSX Edge Compact –> 512MB RAM, 1 vCPU, 500MB vmdk space
      • NSX Edge Large –> 1 GB RAM, 2 vCPU, 500MB vmdk space + 512 MB
      • NSX Edge Quad Large –> 1 GB RAM, 4 vCPU, 500MB vmdk space + 512 MB
      • NSX Edge X-Large –> 8 GB RAM, 6 vCPU, 500MB vmdk space + 2 GB
      • Guest introspection –> 1 GB RAM, 2vCPU, 4 GB vmdk space
      • NSX Data Security –> 512 MB RAM, 1 vCPU, 6 GB per ESXi Host
    • Software
      • Check interop matrix
      • ESXi
      • image
      • vCenter
      • image
      • NSX Manager to participate in a cross-vCenter NSX deployment the following conditions are required:
        • NSX Manager >= 6.2
        • NSX Controller >= 6.2
        • vCenter Server >= 6.0
        • ESXi >= 6.0  (must prepared with NSX vib >= 6.2 )


The component sizing (i.e., small to extra-large Edge) and configuration flexibility in the platform allows adoption of NSX in in across a wide scale of environments.Common factors affecting the sizing and configuration are as follows:

  • The number of hosts in deployment
    • small 3-10
    • medium 10-100
    • large > 100
  • Workload behavior and selection of NSX components mixed with regular workload
  • Multiple vCenter is not the requirements, though offers great flexibility and cross-VC mobility with NSX 6.2 and ESXi 6.x release
  • NSX component placement restrictions depends on vCenter design, collapsed clustering options and other SLAs:
    • Controller must exist in a vCenter where the NSX manager’s registers to
    • Controller should reside in an Edge cluster when a dedicated vCenter is used to manage the compute and edge resources
    • Must consider Edge component placement and properties as described in Edge Design & Deployment Considerations as well the Edge vCPU requirements


Determine how traffic types are handled in a physical infrastructure

  • ESXi Host Traffic Types: Common traffic types of interest include overlay, management, vSphere,vMotion and storage traffic
    • The overlay traffic is a new traffic type that carries all the virtual machine communication and encapsulates it in UDP (VXLAN) frames
    • The other types of traffic are usually leveraged across the overall server infrastructure
    • Different traffic types can be segregated via VLANs, enabling clear separation from an IP addressing standpoint
    • In the vSphere architecture, specific internal interfaces are defined on each ESXi host to source those different types of traffic
    • image
    • VXLAN Traffic: Virtual machines connected to one of the VXLAN-based logical L2 networks use this traffic type to communicate
    • Management Traffic: is sourced and terminated by the management VMkernel interface on the host
    • vMotion Traffic: vSphere vMotion VMkernel interface on each host is used to move this virtual machine state. On a 10GbE NIC, eight simultaneous vSphere vMotion migrations can be performed.
    • Storage Traffic: A VMkernel interface is used to provide features such as shared or non-directly attached storage –> NAS and iSCSI


Determine use cases for available virtual architectures

NSX Primary Use Cases:

  • Security –> Micro-segmentation, DMZ end2end and End-user security
  • Automation –> IT Automation, Developer cloud. Multi-tenant infrastructure
  • Application continuity –> Disaster Recovery Hybrid Cloud and Metro-Pooling


Describe ESXi host vmnic requirements


Differentiate virtual to physical switch connection methods


Compare and contrast VMkernel networking scenarios



   Send article as PDF