Distributed (multiple) hosts / application-level high availability configuration

In a distributed (multi-server) host configuration, AVEVA System Platform components are distributed across multiple VMs located on multiple servers. This configuration is ideal for large-scale and/or highly available (HA) deployments.

Using two or more hosts mostly eliminates the chance that a single failure will disrupt operations. If the systems are located in the same server room or even in the same geographical location, black swan events such as natural disasters, building fire, etc. can still result in a system failure.

To implement a multiple host system, use rack or blade servers. The basic installation can be similar to single server installation, but replicated across two servers. Or, you can split up the System Platform components across multiple servers, and mirror the components, such as Application Object Servers and the Historian Server that require redundancy.

As in the case of a single host system, you can configure multiple VMs on each server, each hosting a different System Platform component,

Redundant components should be installed in separate racks to minimize the chance of a single point of failure.

HIST1 and HIST2: Redundant Historian servers
VIS1 and VIS2: Visualization nodes
AOS1 and AOS2: Redundant Application Object servers
GR: Galaxy Repository and development node with IDE. SQL Server is required for the GR.
VM Host: Virtualization layer

Using multiple hosts provides the following advantages over a single host system:

• Chances of a single point of failure are minimized.

• Easier to scale to accommodate system growth.

• Improved hardware redundancy along with AVEVA's native application-level redundancy.

Multiple hosts do pose some challenges when compared to a single host system:

• They are more complicated to set up and configure.

• They are more complicated to maintain.

• Multiple hosts are more expensive.