It is all good to have the storage, the hosts, the vCenter, etc. We ought to get all this components to work together and to do so, we need to connect them.

Within the VMware world, the basic networking sits on top of Virtual Switches and Port-groups. Without going into too much detail, see below a reminder of what is really important:

  • The Virtual Switch can be associated to a physical switch; similarly, it will have one port (standard virtual switch) or more (distributed virtual switch) uplink ports.
  • A virtual-switch will have at least one port-group attached; these are basically, inbound links. Each port-group will also be associated to either a trunk or specific vlan configuration
  • A Kernel port is a special type of port which will always have an IP address attached to it. The VM Kernel Port is used for management, vMotion and Fault-Tolerance traffic
I will go later through different configuration options as well as to converting a standard switch into a distributed switch.

From the design point of view, as well as applying best practices, in a clustered environment we need to have at least five dedicated networks:

  1. STORAGE – this network will be shared between the Storage (FreeNAS VMs) and the ESXi Hosts.
  2. MANAGEMENT – I will be managing my entire infrastructure over a flat 192.168.1.0 /24 network – this in fact is my local LAN. Furthermore, since this is a virtual network I will need a breakout link into my physical LAN.
  3. VMOTION is a feature provided over VM Kernel Ports, providing host resiliency within a clustered ESXi environtment; it will therefore be shared between the virtual ESXi hosts only
  4. FAULT TOLERANCE – at last, this feature allows seemless failover between VMs. The concept is similar to the ESXi clustering, only applied to VMs. A dedicated network is highly recommended for this feature; it also uses VM Kernel Ports
  5. VIRTUAL MACHINES – this is the network hosting the Virtual Machines hosted by the ESXi host

So here is a diagram:

vcap-lab-networking-generic

A lot of lines … so let me explain briefly…

I am trying to show above, to which network each component connects. For instance, we can see that all ESXi hosts, all the storage nodes and all the other auxiliary components (supporting services) have one link connecting into the MANAGEMENT-1 network and another link connecting into the MANAGEMENT-2 network. Here, by “links” I mean, logical connections – for example, it could be a connection over a physical NIC, over a sub-interface, an SVI, an ether channel, etc…

Furthermore, all ESXi hosts would share the Virtual Machines network – this is where the actual hosted VMs connect to.

And so on …

Hopefully, this will become clearer as we map this diagram to our Layer1 and Layer2 topologies.

Go back to Index

 


Thank you,
Signature
View Rafael A Couto Cabral's profile on LinkedIn



Leave a reply

<a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong>