Distributed (multi-host) network architecture

This is a reference network architecture where System Platform VMs are hosted across multiple hypervisors (host servers).

Recommendations

• Use separate Network Interface Cards (NIC) for IT and OT network traffic.

• Use an external virtual switch and static IP or use DHCP for the primary OT network.

• For the RMC (redundancy management channel) network between redundant Application Object Server (AOS) nodes:

  • Use a separate external virtual switch. This conforms with the earlier recommendation that redundant AOS nodes should be hosted on separate hypervisors (server hosts).

  • Use static IP addresses between the redundant AOS nodes. Multiple pairs of AOS VMs can use the same switch.

• NIC binding order must be configured for the VMs. For details, see Configuring Multiple NICs in the Application Server help.

Network redundancy

Network redundancy ensures high availability and resilience. In a distributed hypervisor (multi-host) environment, network redundancy refers to the design and configuration of the network infrastructure in such a way that no single point of failure can interrupt virtual machine (VM) connectivity or the hypervisor's control and data communication.

There are various key components of network redundancy in a distributed hypervisor environment that are beyond the scope of this discussion of System Platform. Examples include:

• Redundant physical NICs (Network Interface Cards)

• Redundant switch paths

• Network teaming

• Redundant management networks

• Redundant VM network paths

• Storage network redundancy

Network redundancy in a distributed hypervisor environment is primarily implemented at the network layer (Layer 2/3) and not at the System Platform layer (application layer).