Oracle says it will leapfrog competitors in manufacturing intelligence

By now it seems ancient history, but SAP's acquisition of Lighthammer in 2005 marked a sea change in attitudes toward integrating plant operations with the business enterprise.

Lighthammer was one of the first software vendors to develop a solution for enterprise manufacturing intelligence (EMI). These solutions aggregate data from diverse plant assets and, once combined with enterprise transactional information, can track key performance indicators (KPI) and support decision-making related to quality, asset management, or other key resources and constraints.

SAP's arch-rival, Oracle Corp., says it will deliver its EMI product, now called Oracle Manufacturing Hub, “sometime” in 2008.

At the recently held Oracle Manufacturing Summit, the company said the EMI solution will enable these specific actions:

• Contextualize plant-floor data and synchronize it with ERP;
• Provide real-time intelligence for plant operations; and
• Enable a next-generation operations architecture.

While further details about Oracle Manufacturing Hub aren't widely available yet, Alison Smith, a senior analyst with Boston-based AMR Research,believes Oracle is “well positioned to cash in on its database acumen and enterprise experience. ...Between Oracle Daily Business Intelligence, Sensor Edge Server, Business Intelligence Enterprise Edition, and Fusion SOA middleware components, Oracle already possesses the ingredients needed to bring a comprehensive EMI framework to plants.”

Amit Singh, Oracle director of product strategy, says, “We're leveraging all the technology, pulling it together as a solution, making it extensible. Oracle's advantage is we understand the data model top to bottom.”

The Lighthammer product has evolved into SAP MII—formerly xMII, a composite Manufacturing Integration and Intelligence application.

Oracle is poised to launch its solution for enterprise manufacturing intelligence, called Oracle Manufacturing Hub. Note that the system resides "below" the manufacturing execution system (MES).

Smith says while SAP's foray into the EMI space has been highly visible and successful, SAP MII doesn't cover everything manufacturers will eventually need for demand-driven operations. SAP's solution “is not a substitute for an operational data store or historian, as it does not provide a manufacturing data model or tools for manufacturing master data management,” she says.

Without a data repository, adds Smith, there is “no basis for analytics, root-cause analysis, or discovery techniques like ad hoc correlative analysis and inferential model development.” It's important, she stresses, that manufacturers realize the greatest return on EMI investments may come when they “take an architectural approach that blends … different applications to take full advantage of their respective strengths.”

In many SAP MII installations, a data historian acts as the data store. For example, Dow Corning standardizes across plants with its single instance of SAP ERP and plant use of SAP MII and the OSIsoft PI System data historian. In complex, discrete environments, SAP MES by Visiprise can add that same critical data model component.

“SAP Perfect Plant is a not a product, but a portfolio of products ringed by delivery partners, ISVs, and system integrators,” says Paul Boris, SAP senior director, industry solutions. “We didn't fully realize initially how broad an impact the Lighthammer product would have. But it enables advanced technology to be rolled into the portfolio while leveraging a rich legacy.”

Oracle Manufacturing Hub will connect to plant-floor outputs either through historians or directly.

“It provides linkage, comprehensive data model, rules engine, and connections to higher-level enterprise systems,“ says Singh. “The hub collects, cleanses, aggregates, contextualizes, and normalizes data—which aids in the analytics.”

“Oracle is doing a cool thing by bundling a lot of the critical pieces together,” says Smith. “It is providing a piece of the architecture many are struggling to build on their own.”

At end of day EMI is about understanding how operational decisions impact revenue, earnings, and profitability—i.e., performance management—and how the stated goals of the business enterprise are systematically planned and executed—i.e., production management.

The technology challenges involved in normalizing data to allow apples-to-apples comparisons within a business unit or across sites are considerable. A further question is what metrics or KPIs are truly indicative of performance in the most inclusive sense possible.

Manufacturers recognize that they routinely operate without complete information concerning some of their most important operations.

“The master data model is very important at the contextualization level,” says Charles Rieb, VP of product management for process optimization software and services provider AspenTech. “It understands under-the-covers where to get what you're asking for, relating equipment with processes. When you run control data through the model, it's no longer simply temperature, but can be cast as a rate of change”—for example, the beginnings of trend analysis.

This contextualized information also becomes the medium for exposing data to dashboards or higher-level systems, including EMI and corporate business intelligence systems via Web services typically based on the ISA/95 model. Conversely, the data model is a map for deep-dive drill downs when management wants to get at process details behind aggregated information.

To date, says AMR's Smith, manufacturers are falling short in their efforts to brew enterprise intelligence out of production data.

“Many companies are using the technology in a limited fashion to manage tactical KPIs like OEE and packaging line performance, where it's easiest to calculate ROI,” she says. “To a large extent the technology has outrun the market.”

There are companies that have engineered multi-layered architectures, combining technologies to achieve real business intelligence. Chevron, Saudi Aramco, General Mills, DuPont, and Snap-on Tools are said to be among those involved in such efforts.

Kenosha, Wis.-based Snap-on, for example, “is interested in being able to look at an order spike on Thursday, and determine what steps to take now for when that demand hits the plant next Wednesday,” given current production capacity and constraints among all plants, says Smith.

“Manufacturers need to move from real-time performance KPIs to more of a business intelligence mind-set, what we term 'operational intelligence'—business intelligence thinking converging with real-time data,” Smith asserts. “There's confusion about what's needed,” she adds. “It's architecture—not an application—and the architecture must be built in layers.”