Creating a virtual plant

Plant operators are today faced with an ever-increasing number of servers that are used to perform a range of functions within a facility. Each one of those servers has an incremental burden around things such as maintenance, space, power and cooling.

What plant operators strive for is a reduction in the number of servers while still retaining reliability, and with each server that is removed there is a decrease in maintenance and operational costs.

In addition to the sheer number of servers that they need to maintain, companies have to deal with constant hardware and operating system (OS) upgrades, as well as complex system-management issues.

Enter virtualisation, a technology that impersonates or emulates the various layers of a typical computing environment. Through a process of abstraction, it removes many of the limitations experienced when dealing directly with these layers.

There are many different types of virtualisation and each is characterised by the computing component it impersonates. For example:

• Platform virtualisation emulates a PC;
• Application virtualisation emulates an OS;
• Desktop virtualisation emulates a desktop display.

Although traditional computer OSs provide a level of abstraction, they consequently introduce a tight coupling between the OS and hardware. This makes it difficult to move an OS from one machine to another and the OS expects that it is the only user of the processor.

Today’s ’secondgeneration’ hardware abstraction technologies are able to address many of the shortcomings found in earlier virtualisation techniques

There is also poor isolation between applications within an OS, so that applications are generally separated from one another on different computers. This helps guarantee performance and limits any interoperability issues, and may be done for scope-of-loss reasons as well.

Today’s ’second-generation’ hardware abstraction technologies address many of the shortcomings of earlier virtualisation techniques by more cleanly abstracting the OS from the underlying hardware. Called platform or hardware virtualisation, the technology decouples the computer OS from the underlying hardware upon which it relies. Used in data centres around the world, virtualisation can also be applied to industrial-control systems to deliver four important benefits:

• Reduce the quantity of PC hardware required;
• Lessen the frequency and impact of OS and hardware changes;
• Simplify overall system management;
• Improve availability, reliability and disaster recovery.

As server hardware vendors are under constant competitive pressure to increase the performance of their machines, the performance improvements in servers generally increase at a faster rate than applications. This sometime results in the under utilisation of machines that potentially could be used for other tasks.

While multiple applications can often be run at the same time within an OS to take advantage of this increasing capacity, it is generally not done for scope-of-loss reasons and to avoid conflicts from a performance or compatibility perspective with other applications.

Deploying virtualisation, however, allows multiple virtual machines, each running their own OS and application, to be operated at the same time on a single physical machine. This is achieved while guaranteeing that a given virtual machine gets exactly the amount of resources required to do its job, and ensuring any issues with one virtual machine won’t impact on another.

This type of consolidation reduces the need for multiple machines and maximises the use of hardware resources.

Industrial plants are, therefore, able to cut down on the number of physical computers they require, which has a direct correlation with associated running costs. With each machine removed, there is a decrease in maintenance and operational expenses.

Virtualisation also allows plants to undertake expansions without adding new hardware. That’s because platform virtualisation can run many virtual machines on a single piece of hardware. Machines can continue to be added as long as there are sufficient resources available to meet their operational needs.

Plant operators and manufacturers consistently struggle with the disconnect between the fast-paced and obsolescence-driven world of hardware and OS manufacturers and that of process control operations, which operators want to remain static.

Overhauls of core assets are extremely expensive for plant managers, who don’t want to be forced into change or upgrades that are not on their own terms. But in an industry where technology has to remain current, plant operators are struggling to keep up with the rate of OS and hardware change they are facing.

Plant managers often don’t see a business benefit in moving forward with OS or hardware upgrades. They are doing so for hardware reliability reasons, along with the ability to receive support and security patches on the operating systems.

None of these items increases a plant’s output, but are merely being done to avoid downtime and ensure support. Operators see OS and hardware upgrades as an unnecessary disruption and burden.

Virtualisation helps reduce the strain on industrial facilities by allowing existing hardware to be maintained for as long as it is able to provide the minimum levels of performance an application requires of a virtual machine.

By sticking with existing hardware, plants can reduce the cost of system upgrades. They can also stay on the same operating system for a longer period of time, since the ability to run on a physical piece of hardware is no longer dependent upon the underlying hardware itself, but rather on what the virtualisation environment can support.

Less disruption
Virtualisation minimises the disruption to plant operations when hardware changes are required by avoiding the need to reinstall an OS. Thanks to more advanced virtualisation features, a computer’s OS and applications don’t need to be shut down during a hardware replacement.

Operations can continue without interruption while this work is taking place.

Virtualisation supports a strategy of ’build once, deploy many’. It enables the hardware configuration, OS and application to be contained in a single ’capsule’. Any new instance of this capsule is exactly the same, thus reducing configuration errors and installation time, and ensuring a more reliable and repeatable result.

With virtualisation, plants can undertake expansions or upgrades without the need to add new nodes to the control system and without having to perform fresh operating system and application installations.

Virtualisation also offers improved diagnostics, allowing operators to monitor system performance and access the desktop of any node. Health and status information can be viewed from a single, integrated userinterface. Plus, it simplifies remote management. Whereas a physical system requires a dedicated remote desktop and management functions, these capabilities come for free by having remote access to the virtualisation console.

The latest virtualisation offerings enable industrial facilities to simplify and automate the key elements of disaster recovery: setting up disaster-recovery plans, testing those plans, executing failover when a control-centre disaster occurs or as the event requires, and failing back to the primary control room. This makes it possible to provide faster, more reliable and more affordable disaster-recovery protection than previously possible.

Virtualisation also helps applications achieve higher availability than they would be able to support natively. For example:

• Snapshots allow processes to roll back to a known point in time instantly;
• Users can employ a common set of more reliable hardware that can be shared among multiple virtual machines;
• If hardware should fail, it can be replaced with dissimilar hardware without impacting the virtual machine.