Skip to main content

VCF 9 Networking : Centralized vs Distributed Connectivity

 

VCF 9 Networking : Centralized vs Distributed Connectivity

With the release of VMware Cloud Foundation 9 (VCF 9.0), VMware has redefined the private cloud networking model by introducing both centralized and distributed connectivity. This dual approach provides flexibility for organizations to choose between traditional edge-based routing and modern, host-level distributed networking.

Why New Networking in VCF 9?

Prior to VCF 9, NSX networking was largely edge-centric—requiring dedicated Edge clusters to handle north-south traffic. This created scaling and operational overheads. VCF 9 introduces a cloud-like networking abstraction with:

  • Native VPCs (Virtual Private Clouds) for tenant isolation.

  • Transit Gateways (TGW) to interconnect VPCs and external networks.

  • Simplified bootstrapping – NSX VIBs are embedded in ESXi for easier enablement.

  • Improved lifecycle and visibility via integrated VCF Operations.

These enhancements align with VMware’s goal of offering public cloud simplicity with private cloud control.

Core Building Blocks

Virtual Private Cloud (VPC)

Each VPC provides isolated networking for workloads. Users can create subnets, security groups, and routing policies.

Transit Gateway (TGW)

TGWs act as the backbone of connectivity, linking multiple VPCs and connecting to external or on-prem networks.

External Connectivity Models

VCF 9 offers two approaches for north-south traffic:

  • Centralized (CTGW) – traffic flows through NSX Edge clusters.

  • Distributed (DTGW / Edgeless) – routing happens at the ESXi host level.



Centralized (CTGW) Connectivity

In the centralized model, external traffic exits via NSX Edge VMs hosting Tier-0 Gateways.

Workflow:

  1. VPC traffic routes to TGW.

  2. TGW connects to Tier-0 gateway on Edge Cluster.

  3. Tier-0 peers (BGP/static) with physical routers.

Advantages:

  • Centralized control and policy enforcement.

  • Ideal for complex routing and multi-WAN setups.

  • Easier integration with existing NSX Edge-based services (NAT, LB).

Limitations:

  • Higher latency (additional Edge hop).

  • Resource overhead of Edge clusters.

  • Possible bottlenecks under heavy traffic.

Illustration:



Distributed (DTGW / Edgeless) Connectivity

The distributed model removes the dependency on Edge VMs. Each ESXi host becomes capable of directly routing external traffic.

Workflow:

  1. TGW maps to a VLAN connected to the physical network.

  2. Each host handles north-south routing directly.

  3. Network services (NAT, firewall) are distributed at the host level.

Advantages:

  • Lower latency and improved performance.

  • No need for dedicated Edge nodes.

  • Simpler and more scalable for lightweight or edge deployments.

Limitations:

  • Less centralized visibility.

  • Some advanced services may still need Edge nodes.


Comparison Table





End-to-End Flow Example

  1. Project Admin defines a Transit Gateway (TGW).

  2. VPC Admin creates VPCs and subnets.

  3. VPC traffic flows to TGW.

  4. TGW connects to either:

    • Tier-0 Gateway (centralized)

    • VLAN uplink (distributed)

  5. External IPs and BGP peering complete the setup.

Visual Flow:

VCF9 Networking Overview

When to Choose Each Model

Hybrid adoption is also supported—some domains can run centralized edges while others leverage distributed exits.

Extract:

  • VCF 9 unifies NSX networking into a modern, VPC-driven model.

  • Distributed networking simplifies infrastructure for smaller or edge environments.

  • Centralized networking remains relevant for policy-rich or complex routing needs.

  • Together, they deliver flexibility, scalability, and cloud-like networking agility.

Resources good to have on this

Thanks for the reading . Happy learning 😊

Comments

Post a Comment

Popular posts from this blog

Changing the FQDN of the vCenter appliance (VCSA)

This article states how to change the system name or the FQDN of the vCenter appliance 6.x You may not find any way to change the FQDN from the vCenter GUI either from VAMI page of from webclient as the option to change the hostname always be greyed out. Now the option left is from the command line of VCSA appliance. Below steps will make it possible to change the FQDN of the VCSA from the command line. Access the VCSA from console or from Putty session. Login with root permission Use above command in the command prompt of VCSA : /opt/vmware/share/vami/vami_config_net Opt for option 3 (Hostname) Change the hostname to new name Reboot the VCSA appliance.   After reboot you will be successfully manage to change the FQDN of the VCSA . Note: Above step is unsupported by VMware and may impact your SSL certificate and face problem while logging to vSphere Web Client. If you are using self-signed certificate, you can regenerate the certificate with...
2 comments
Read more

High-Speed Networking for VMs: VMXNET3 Now Supports Up to 65 Gbps

The VMXNET3 adapter is VMware's high-performance virtual network adapter, designed for optimized performance in vSphere environments. It is the default and recommended adapter for most modern workloads on ESXi , delivering high throughput, low latency, and support for advanced networking features. With the release of vSphere ESXi 8.0 Update 2 , VMXNET3 introduces new capabilities that significantly enhance virtual network performance. What’s New in ESXi 8.0 Update 2 for VMXNET3? Traditionally, VMXNET3 reports a default link speed of 10 Gbps to the guest operating system. However, the actual achievable throughput often exceeds this value—thanks to technologies such as multi-queue support , Receive Side Scaling (RSS) , and Large Packet Offload . Starting with ESXi 8.0 Update 2 , this limitation can now be overridden , and VMXNET3 can be configured to support custom link speeds of up to 65 Gbps . This enhancement enables improved flexibility and throughput for high-performance w...
Post a Comment
Read more