Service-oriented architecture (SOA) projects are sometimes most in peril just as they are being created. They bog down after the first meeting. That's the one in which the business, development and data sides get together to decide "what a customer is." Deadlocked on that decision, the parties agree to form a committee—you can guess the rest.
To address these issues, master data management (MDM) is emerging. It is, among other things, a discipline to ensure that critical projects such as BI and SOA are done right. MDM ensures that data definitions are consistent across the organization so BI systems, for example, display accurate data. As business process management (BPM) rises to the fore in the face of economic recession, it too, enforces a need for MDM capabilities.
"All applications need a single version of the truth—consistency," says Wayne Eckerson, research director at The Data Warehouse Institute (TDWI) in Renton, Wash. "How do you do it?
"You could do it out of the data warehouse, but it was never really designed for that," he says. Alternatively, he adds, "you could build an MDM and have any applications, including the data warehouse, pull from that.
"MDM fits with SOA. It is a data warehouse for the rest of the enterprise. The application that it supports is decision making and planning, but mostly decision making." Meanwhile, as BPM and business rules engines spread, decision making takes on more of the aura of operations—sometimes 'real-time' ops.
"If we decide that we want to have a message that flows through an organization that encapsulates customer data … but [the systems] don't agree on what a customer means, then there is a high likelihood they are going to misinterpret the data," says Marty Moseley, chief technology officer at Chicago-based data integration tool maker Initiate Systems Inc.
Moseley spoke in a recent Initiate Systems podcast on MDM and SOA. SOA without MDM can help you make "big mistakes in record time," Moseley, jibing, suggests.
Increasingly, organizations are designing SOA into the MDM architecture from the beginning, says Dan Power, president and founder of consulting firm Hub Solution Designs Inc. in Hingham, Mass. This creates challenges in meshing the real-time realities with the need to keep the data accurate. To do this, organizations are building in business rules engines to monitor the data, he says.
"The main issue is establishing an MDM strategy that ensures that business services are able to get at the right data that is trusted," says Judith Hurwitz, president of research and consulting firm Hurwitz & Associates in Newton,Mass. TDWI's Eckerson explains, "Semantics are the big issue. It is easy to wire applications together, but getting companies or parts of companies to agree on what the messages really are—what is a customer or what is a sale—that can be hard."
SOA and MDM: GIGO 2.0?A five part special report
Part 1: SOA with MDM prevents messaging confusion
Part 2: SOA and MDM: New techniques address old problems
Part 3: Data architecture project practices with SOA and MDM
Part 4: MDM brings SOA and BPM closer together
Part 5: With MDM and BPEL, business users become data stewards