On top of all of this, virtualized infrastructures can self-heal. If one physical server goes down, for example, the virtual system can automatically restart the lost applications on other physical servers to quickly get production running again or even prevent it from stopping. Hardware failures no longer need to be major production-halting events.
Design Considerations
From a hardware standpoint, a virtualized infrastructure in a manufacturing environment typically requires two to four physical servers with sufficient memory to host all of the virtual machines, enough disks to run a plant's applications at the required speed, and switches and cabling.
The process of transitioning to a virtualized system should begin with an audit to assess design specifications and business objectives, and then to identify the functional and informational requirements for the virtualized infrastructure. Some questions that will need to be addressed at this stage include:
- How much RAM, CPU and disk I/O do your applications require?
- How many client workstations will be deployed in the virtual environment?
- How many servers will be needed?
- What kind of network switching will be used?
And while understanding how a virtualized infrastructure will support current operations is an obvious immediate need, it’s also critical to be thinking ahead in order to anticipate future needs. Don’t short change what operations might be doing in the next five years. Systems grow and evolve over time, and designing room for growth into a virtualized infrastructure will allow for greater agility, making it easier to deploy new applications down the road.
Some of these design considerations for future growth could include ensuring enough switch ports and communication throughput to add a third or fourth server at a later point, or having the capacity to add memory to the servers at a future date.
To Build or to Bundle?
When procuring hardware for a virtualized infrastructure, there are two options.
The first option is to build the infrastructure from scratch, which requires ordering all of the necessary equipment, assembling it and commissioning it. This can be a burdensome and time-consuming approach: Equipment must be ordered from multiple vendors and system design, fabrication and testing can take weeks. There also are added costs of hiring certified installation professionals or trained technicians to provide support.
The alternative to this piecemeal approach is a bundled solution. Bundled offerings are preassembled systems that include all of the hardware, software and documentation for a virtualized system in one turn-key solution. These solutions are assembled using industry best practices in areas like cable management, system grounding and labeling and have been pre-engineered so that the infrastructure design effectively addresses all system needs.
Bundled offerings typically include implementation services to execute the on-site network configuration and integration. From design to deployment, a virtualized system can be up and running in a matter of days for a bundled solution versus a matter of weeks for a non-bundled solution.
When going the bundled route, it's crucial to understand that not all solutions are designed to exclusively to address the unique characteristics of a manufacturing environment. This is an important consideration to keep in mind, because different industries have different priorities when it comes to downtime, system complexity and cost.
A five-minute system crash in a corporate office environment, for example, could lead to the temporary loss of email and business systems. An inconvenient and perhaps even costly event. But a five-minute server crash in the manufacturing environment can lead to a catastrophic downtime event, such as the loss of a high-value batch.
Similarly, some bundled solutions that are designed for environments such as corporate data centers could include features that are in excess of or not relevant to the needs of a manufacturer, which can drive up costs. Consider whether the solution you choose is purpose-built and purpose-priced to meet your unique needs.
Simplify Your Support
One of the appeals of a virtualized infrastructure versus traditional client-server architecture is simplification of ongoing system management. A virtualized system makes it possible to control and manage operator and engineering workstations from a central location. But as with anything in a production facility, it will still require servicing and support for maintenance, repairs and upgrades.
Selecting a solution provider that offers technical support with their product can make life easier for the maintenance manager – they'll have one phone number to call for whatever questions or support issues arise across the system’s life cycle.
Also consider what additional levels of support may elevate the organization's performance given new capabilities enabled by virtualization. A virtualized environment simplifies remote monitoring, for example, allowing a provider to monitor the complete virtualized infrastructure, identifying and troubleshooting issues, or contacting maintenance personnel immediately to alert them of a problem – all from an off-site location. This is particularly important for manufacturers that don’t have an IT administrator in their plants or lack the expertise needed to maintain and service the virtual infrastructure.