Real time enterprises held back by disparate network architectures

2 mins read

Disparate network architectures at the factory and business levels are set to become one of the biggest barriers to more automated, real time supply chain operations for manufacturers. Brian Tinham reports

Disparate network architectures at the factory and business levels are set to become one of the biggest barriers to more automated, real time supply chain operations for manufacturers. At the plant level, although more are going towards the Microsoft .Net architecture, according to Dan Miklovic, senior consultant with analyst Gartner, 95% remain proprietary – and reality is probably nearer 99%. Meanwhile, at the enterprise level its mainstream Websphere, Java, Windows 2000, Linux, etc. “There’s a huge disconnect,” says Miklovic. And while that’s always been the case, as companies pursue the ‘real time enterprise’, customer-driven goal and find themselves having to get away from batch transactions to get more agile, it is becoming a real issue. Manufacturers that have long since stopped being vertically integrated, relying instead on contractors and supply chains woven together with JIT (just in time) agreements, are in the thick of it already. But it’s going to get worse as faster communication and better linking from top floor to shop floor are demanded. Today, for most information flows, capacity planning and scheduling systems talk plant to plant with ‘latest data’. At best, shop floor systems link up from the plant into ERP, sending data out into the supply chain to other ERP systems and back down to plant systems – and then all the way back, and so on. “We need manufacturing systems, machines and devices to be able to talk to each other direct across enterprises,” observes Miklovic. But this isn’t trivial. Either, manufacturers bite the bullet and normalise network architectures across the levels – not favourite and dubious in terms of technical, secure feasibility – or they pay what it costs to integrate, which is costly. The third way is peer-to-peer communications, and despite being entirely possible, and in action already with web technologies, this is bedevilled by security, privacy, management and bandwidth issues. Although network policy systems can help deal with otherwise chaotic systems, this sounds futuristic – and for now it is. If fieldbus (digital plant communications) standardisation took around 15 years, how long is this going to take? No one will give you a sensible estimate, but the clue may yet be in the Web services base standards arena. In which case, realistic systems at realistic prices could be less than five years away. For now, Miklovic concedes that where this level of real time networks are concerned, “only 10% are doing anything, the remaining 90% are just thinking about it. And they’re saying ‘we know we’ve got to go there, but we’re not going to do so until our clients force us and, guess what, they’re not forcing us’.”