In his 1995 article, Robert Brooks discussed integration. In particular, he explored the relative merits of open and proprietary systems, concluding that, at the time, the proprietary approach still had much to recommend it. Have his views changed?
Some things have changed hardly at all since I wrote my article in 1995. It's still reasonable to define integration as linking two or more pieces of equipment together in ways that bring benefits to the user. And the objectives of integration are also still much the same: reducing costs. Systems that are more flexible and functional and which are easier to commission and then maintain through an effective diagnostics capability, all help to reduce the costs of ownership.
To achieve this, users are still looking for open connectivity, so that they can pick the 'best of breed' product from a range of suppliers, without having to worry about compatibility issues. In reality, ten years on, are they still typically only trusting products from one vendor when true integration is their primary objective, and has technology really moved forward enough to challenge this strategy?
If the basic ideas and objectives have remained unaltered over a decade, a lot has changed. For example, my original article discussed the use of Ethernet as an integration technology and concluded that it was often unsatisfactory, particularly for the control of drives, because of the unpredictability of message transit time. Today, with Ethernet links routinely operating at 100Mbit/s, utilising full duplex and with managed switches, these problems have been largely overcome. For example, messages are now directed only at the desired node via managed switches, greatly improving the throughput.
Ethernet technologies have subsequently gained ground in industrial automation schemes and now form the basis of several new developments on the industrial networking front, including Modbus on Ethernet, Omron FINS on Ethernet and Ethernet Powerlink. Also, the Common Industrial Protocol (CIP) implemented over Ethernet has led to Ethernet/IP (Industrial Protocol) which will standardise communications over Ethernet as Profibus and DeviceNet have for done for Fieldbus networks. These developments give users much more choice in networking solutions.
Other big changes over the decade have been in the areas of data handling and in communications with higher level systems. Today's automation installations are required to link directly with factory management and accounting systems. ActiveX components are one of the tools, which help with implementing links between the production automation and commercial levels of a business.
Going back to that important issue of the user being able to choose 'best of breed' products, one of the more recent developments is FDT/DTM, or Field Device Tool/Device Type Manager. This technology is led by a joint interest group set up in 2002, which includes many of the leading automation suppliers. FDT/DTM is not concerned with connectivity hardware - rather it is a tool intended to improve the user configuration environment, facilitating straightforward integration. It is one of the modern faces of openness, which is still high on the wish-list of automation system users.
To explain the functioning of FDT/DTM, it is easiest to think of FDT as a receptacle and DTM as the link to that receptacle. Where a spreadsheet or Visual Basic program may be a receptacle for ActiveX control, the FDT receptacle will more usually be a specially written piece of software provided by automation vendors.
DTM is a file, which not only allows configuration of a product, such as a drive, to a Profibus network (the limit ten years ago), but now also offers the ability to monitor, visualise and even have parameters uploaded/downloaded from the FDT environment, via the network. This includes any vendor's product, as any automation vendor can supply DTM files for their products. The result is a truly open system, with easy access to all data.
Compared with a decade ago, there is now much more emphasis on the software environment for integration. Users are increasingly unhappy about having to work with numerous different software packages. They want a single software environment, irrespective of network type or equipment supplier. They also want to access all devices from a single connection on the network. FDT/DTM is an important step to meeting these requirements.
Even ten years ago, users could achieve effective integration, but at a price - the lack of an even remotely uniform environment meant that the time and, therefore, cost overheads in configuring and maintaining an integrated installation were high. A good compromise, as I said in my earlier article, was to buy all of the equipment from a single vendor, but this made it difficult, if not impossible, to make sure that every product was the 'best of breed'.
Today, the situation has substantially improved, and the gap between user requirements and vendor solutions has diminished. Further improvements are imminent, and the outcome of these developments will be crucial for all concerned. Let's see where they lead us in the next ten years!
Robert Brooks is with Omron Electronics