Last week DELMIA Quintiq R&D released a new High Availability white paper. This "best practice" white paper describes the capabilities of the software and software configurations to ensure high availability. In this post, I would like to highlight some key parts.

Business critical operations require high availability of supporting software services as unavailability of those services may directly impact production. When designing software services architecture, it is therefore essential to consider the worst-case scenario that could occur and the maximum recovery time that results from it.

Availability is expressed as the uptime of a service over a period of time (typically a year) expressed as a percentage. For example, a required (yearly) availability of 99.0% means a maximum downtime of 3.65 days per year. The DELMIA Quintiq software can technically support availability up to 999.999% meaning only 957 seconds per year of downtime. This however requires a careful design of the hardware configuration, software configuration and orchestration.

Depending on the required availability, DELMIA Quintiq supports different fail-over architectures ranging from simple to advanced.

Simple HA Architecture

In this scenario, fail-over is managed by the VMware tools in combination with periodic backups. A virtual machine (VM) is restarted after being moved to another VM node. This scenario is easy to configure and does not require a special configuration of the DELMIA Quintiq software.

The diagram below describes this scenario in more detail.

Advanced HA Architecture

In the advanced scenario, a second standby DELMIA Quintiq environment runs without datasets loaded. When the primary environment fails, the only thing needed to recover is loading the datasets in the secondary environment which can be performed really fast (GB/Sec). The diagram below describes this scenario in more detail.

It may be clear that there are quite some prerequisites to implementing an advanced fail-over architecture. For example, high hardware requirements, data-set store mirroring, orchestration, load-balancing configuration etc. requires careful design, implementation and testing.

See the whitepaper for more details about the software's capabilities and HA configuration. Looking forward reading your comments.