Here we go again.
Hoping to create a lock on the manufacturing software arena for its products, Microsoft formally unveiled Windows Distributed interNet Applications for Manufacturing last week. DNA-M, as it is being acronymed, is a "technical architecture" for building applications and linking data systems from the factory floor to the corporate databases. It represents the forward momentum Microsoft has generated with Windows NT in business applications, as well as the interoperability efforts that can be traced back in time to OLE for Process Control, OLE, and DDE. And before that, industry watchers can recall the Manufacturing Automation Protocol (MAP) that came out of the automotive world in the mid-1980s.
It is worth noting that, like many of the more far-reaching software announcements in industry today, DNA is not a product, but a strategy. The strategy clearly is set up for Microsoft's benefit, but if it delivers on its initial promise, and if the user community is active enough in its support, it will also clearly benefit the users.
"Integration is a major key to the success of deploying a modern manufacturing system, but piecing the components together to produce such a system is often difficult, time-consuming and expensive," said Steve Ballmer, president of Microsoft, at a Feb. 23-24 industry meeting in Seattle, WA. "We're striving to make this process easier and more cost-effective for customers by developing this . . . new architecture, which will help combine the Windows platform, various leading line-of-business applications and legacy solutions to create manufacturing-specific Digital Nervous Systems."
Windows DNA for Manufacturing is being developed and supported by industry-leading developers, hardware vendors and systems integrators. These include (according to Microsoft's carefully alphabetized list) Aspen Technology, Camstar Systems Inc., Cincom Systems Inc., Compaq Computer Corp., Ernst & Young, Honeywell Inc., Iconics Inc., Intellution Inc., Macola Software, Marcam Solutions Inc., National Instruments, the OLE for Process Control (OPC) consortium, Rockwell Automation, SAP AG, Sequencia Corp., Siemens AG, Symix Systems Inc., USDATA Corp. and <%=company%> . Other companies, such as OSI Software, Inc., Fisher-Rosemount, Baan and Phoenix Contact have, or are developing, DNA applications. Whether any company with any Windows NT-compatible software can claim to be part of the DNA bandwagon remains to be seen.
However, the potential seeds of discord are already in place for these third-party vendors. Many of them have built up their own integration toolkit or architecture-and all of them purport to offer wall-to-wall compatibility for Windows-based software. Consider these architecture "brandnames:" Plantelligence (Aspen), PlantWeb (Fisher-Rosemount), iCore (Intellution), SAP Business Framework (SAP), FactorySuite (Wonderware's flagship product and integration architecture) and EnterpriseBatch Production Manager, Sequencia's integration path. Many of these companies are also key players in the OPC Foundation (which originated as OLE for Process Control, but continues to evolve into a larger scope as OLE itself does).
COM and Internet
According to a Microsoft white paper published last fall (Windows DNA: Building Windows Applications for the Internet Age, http://www.microsoft.com/dna/overview/dnawp.asp#web), a wide range of Microsoft products, currently under development, will be the embodiment of DNA-M. Two of the more noteworthy ones are DCOM, Microsoft's preferred method of building software components, and Internet Information Server (IIS), Microsoft's method of providing Internet capabilities within a corporate intranet (or, for that matter, within a distributed process control system).
DCOM lines up pretty clearly against Corba, the object-oriented rule set in Unix systems. In theory, either DCOM or Corba should be capable of allowing Unix-to-Windows or Windows-to-Unix communications. But in reality, Microsoft would seem to prefer that everyone use DCOM for that purpose. Most of the more Microsoft-centric control vendors have complied, licensing Visual Basic for Applications (VBA) as part of their own product suites, and seeking certification for a "designed for Windows 98 or NT" seal from a third party authorized by Microsoft.
The real fun begins in looking at the Internet Information Server. Microsoft is planning four levels of Internet integration for parties seeking to use Internet standards for factory automation:
The upshot of this is that while it will be technically feasible to link, say, controller setpoints on the plant floor to an Excel spreadsheet in the corporate HQ miles away with a browser-neutral format, Microsoft, and the vendors aligning themselves with it, will go for the Internet Enhanced level of integration. Bye-bye Netscape.
Although Microsoft has made enormous strides in the manufacturing marketplace with Windows NT, the battle is far from over. Microsoft's Ballmer, in making the DNA-M presentation, cited market statistics from AMR Research (Boston, MA) to the effect that in 2002, 66% of all ERP software will be shipped on the NT platform. A two-thirds market share in a market where Microsoft had only a vestigal presence three or four years ago is a huge success, of course. But it is significantly behind the over-90% market share that Microsoft has in office automation generally.
And office automation is, of course, the major-league playing field that Microsoft wants to dominate. Pat Kennedy, president of OSI Software, is fond of pointing out that while there are potentially millions of PC users in manufacturing, there are already hundreds of millions of users in office automation. OSI has been Microsoft-centric for years, and already has built DNA-M applications using IIS.
All this would seem to be a healthy evolution of the manufacturing automation world, which has suffered the islands-of-automation issue since its inception. However, it also makes the leading developers of manufacturing automation systems wholly dependent on Microsoft, and its development schedule for NT 5.0, now seriously delayed. The tug of war between operating-system development, applications development and systems integration will continue.