Monday, December 29, 2003		Guest Speak

APPLICATION of innovative technology in large-scale manufacturing process has been one of the key elements when we talk of reducing cost in terms of operation as well as overheads. During the last decade, developing and designing a typical automation software application was relatively simple in terms of concept. Today it has become more complex in terms of integration with various processes, complete software architecture and its operating platform.

In the past, there have been many classes of servers that provided "services" for industrial deployment, but they were usually limited to singular functionality. There were so-called "tag" servers that collected data from plant floor devices, scaled it, checked for alarms and events, and then distributed it to client applications. There were communications servers, which simply managed communications between industrial applications and distributed plant floor equipment. There were calculation engines that processed data on-the-fly as a production process was running, for reuse elsewhere within the ongoing process.

Programmable logic controllers (PLCs) and dedicated process controllers ran equipment or processes on loops. Plant floor sensors, actuators and recorder devices were interfaced to these control platforms. Human-machine interface (HMI) software provided the PC-based visualisation needed to interact with the process under control. The problem was as engineers built more functionality and sophistication into their applications, they were adding more data tags, a lot more scripting and many more alarms. In addition, to scale systems up to expand or enhance production lines meant they had to add more PCs, PLCs and control devices. This meant networks needed to grow as well, and they were usually segmented so that throughput would not be impacted. While system engineers borrowed heavily from the client/server technology deployed in the business world, they could not really use it efficiently because of the difference in the nature of the business and industrial worlds.

Application servers

As applications grew bigger and more complex, new problems were created for administering networks, managing application changes, scaling applications to add new lines or enhance application functionality and efficiency. Change management has become an increasingly large cost element in re-engineering. The only realistic solution today is to borrow a page from the business IT world and deploy application servers.

As a matter of the typical historical development pattern, the industrial world has lagged the enterprise world by several years in its adoption of IT technologies and adaptation of them to the factory floor. This has meant that application servers were most often used for business applications, serving up application modules and database information for use in customer resource management, e-business, financials, human resources and other enterprise applications. The problem is that most application servers provide services in a transaction-based environment. And that doesn’t work in an event-driven industrial environment because the ground rules are different in a factory than they are in an office.

The requirement for a true industrial application server has come about because of the need to simplify many automation applications. As an example, many users of human-machine interface (HMI) applications customise their applications by building in scripting to perform certain tasks. The problem that has resulted from the heavy use of scripting, however, is that it has turned seemingly ordinary HMI applications into huge and complex ones that are very difficult to maintain and enhance. They’re also difficult to scale for expanding plant floor operations. Removing the application logic from scripting and incorporating it in the "factory" business logic managed by an application server can "thin down" these applications so they perform better, are more scalable and are easier to administer and maintain.

An Industrial Application Server (IAS) is like nothing the IT world has seen before. It looks and acts like a traditional application server, but it provides a whole new layer of real-time data acquisition, event management, data manipulation services and collaborative engineering capabilities that have been designed from the ground up for use in industrial environments. It provides a new tier of real-time data acquisition, alarm and event management, data manipulation services and collaborative engineering capabilities that have been designed from the ground up for use in industrial automation applications.

Expectations

The Industrial Application Server is an infrastructure for simplifying the development, deployment, maintenance and administration of distributed automation applications.

An industrial application server performs the same basic functionality as its business counterparts, which is, providing services and data to multiple applications – but it does so while fulfilling unique requirements of the industrial world.

As for example, it must operate in real time to handle millisecond transaction and event speeds, must be able to monitor and respond to extremely high volumes of asynchronous data and event messages (thousands of messages per second), must be a peer-to-peer implementation to facilitate interaction with thousands of plant floor devices as well as provide access to applications from multiple sites, both local and remote, must be deterministic, providing the ability for things to be done in a set order. It must facilitate the use of information as part of the process under control because certain events rely on receiving data during the process, not after it’s completed.

Such technology applications empower decision-makers to achieve their business goals, without abandoning prior investments in automation systems, production processes or intellectual property.

However, easier said than done, it took more than just copying standard application server technology to meet the needs of the industrial world.