Want to Buy a Bridge? The Promise of Service-Oriented Architecture

Daniel R. Matlis

Daniel R. Matlis

Medical device companies have been implementing manufacturing information systems for more than 30 years. These systems were often part of automated manufacturing initiatives aimed at increasing product quality and lowering manufacturing costs. For the first time, manufacturers began to roll out computer workstations on the shop floor as part of these automation projects. They went by names such as CIM (computer integrated manufacturing), SCADA (supervisory control and data acquisition) and SPC (statistical process control), among others.

The computer derivative of Parkinson’s Law states “Data expands to fill the space available for storage.” The availability of computers and the data storage they provided led to a major side effect: the creation of “data islands.” Every automated machine with an attached computer began collecting historical information and process data.

However, while computers had reached the shop floor, it would take some time for networks to catch up. Thus, these data islands were connected by “island-hoppers”’ (Figure below), the process by which data are placed into a system by walking over to the next data island in the process and manually entering it.


For many years, device manufacturers have desired the ability to make production information, processes and resources more transparent. Companies have often looked for ways to unlock operational data to empower decision makers—from operators to plant managers and corporate executives—to make informed, timely decisions. This level of integration has eluded companies…until now.

Today, software manufacturers are moving away from proprietary systems and interfaces and are working together to develop open standards. These connectivity standards have lead to the development of a new class of software, known as middleware. This standards-based software enables the collection, analysis and presentation of critical, real-time data to decision makers throughout an organization. Known as Service-Oriented Architecture (SOA), this technology provides a bridge that allows for the connection of data islands in an efficient and effective manner.

How the “Bridge” Works

The SOA concept is based on the use of Web services to support the requirements of software users. In a SOA environment, data islands make resources available to other network participants. The users find that these independent services are accessed in a standardized way.

Unlike traditional system architectures, where software is stacked to meet a particular department’s need, SOA is composed of business functions that are loosely coupled. These functions or services interoperate based on a formal definition independent of their platform and programming language. The interface definition hides vendor proprietary and technology-specific implementation (such as Java and .NET).

Functionally, SOA seeks to transform stand-alone systems and their data islands into “services” that can be accessed with a common interface, regardless of the location or technical makeup of the system. These services must then be provided with the means to communicate with each other. The communication can involve either simple data passing or could involve two or more services coordinating an activity.

For SOA to be effective, these functions must be well defined, self-contained and must not depend on the context or state of other services. Today, Web services are the most likely technology used to help SOAs achieve robust connections (Figure below).


The use of SOA has led to development of standards such as Business Process Execution Language (BPEL), which takes the service concept one step farther by providing a method of defining and supporting workflows and business processes.

Building That Bridge

The implementation of SOA requires a strategic process, not a tactical project. The journey to SOA begins by defining clear and specific business requirements. Business needs—not technology—must drive this process. An IT department within an organization must undergo a significant shift on the need to “align IT with the business” to the reality that IT must support the needs of the business. To this end, IT organizations are looking to provide the tools required to make business decisions, based on real-time information, in a cost-effective manner.

Another key step in the process is to evaluate the current IT portfolio. This step seeks to identify and eliminate applications developed as stopgap measures to address a business need that was not provided by commercial off-the-shelf software. The proliferation of these “pop-up” applications, which are critical to running the business and the release of quality product to the market, has also led to increasing IT operational costs due to the ad-hoc and uncoordinated development associated with “pop-up” applications.

To facilitate the implementation of SOA, software suppliers are moving away from proprietary software interfaces that result in unique applications. While the industry has tolerated proprietary systems in the past, leading software providers not only need to operate and acquire data from the process in real time, but also must provide the capabilities to share and analyze the data to be transformed into actionable business information. Given this need for an integrated environment, next generation IT solutions must be designed to be part of a SOA.

Bridging to the Future

Over the next few years, IT organizations will play an intricate role in technology decisions at all levels within a medical device company. This is evident from the integration of what used to be separate automation and IT organizations into a single organization spanning both disciplines. This will lead to more efficient use of IT resources requiring the integration of systems using SOA and the elimination of redundant systems and data.

Additionally, due to regulatory requirements, medical device manufacturers are looking to define authoritative sources of data across the organization to minimize compliance risks by controlling and providing visibility into the manufacturing process.      

The SOA for automation and manufacturing hinges on the development of an integration and intelligence infrastructure that collapses the traditional “island-hopping” model and transforms the model into a collaborative one.

The use of SOA significantly increases situational awareness in today’s global and outsourced markets. With SOA, today’s OEMs can communicate with contract manufacturers, raw materials suppliers and supply chain managers instantaneously and cost effectively by building bridges between their respective systems of record. This eliminates the need to build custom interfaces with every system and, therefore, minimizes compliance and validation costs.

Although SOA will revolutionize the way companies share information internally—as well as with partners, customers and regulators—we must always keep in mind that technology can only enable change and should never drive it. Change must be rooted in sound business needs and in industry and regulatory requirements.

Once you develop your business requirements, you’ll probably want to buy this bridge.

Daniel R. Matlis is founder and president of Axendia, a research firm providing forward-thinking and actionable strategies focused on the use of technology in the life-sciences and healthcare markets. With more than 15 years’ experience in the life-sciences market, he has been involved in projects spanning R&D, clinical, manufacturing, marketing and information technology. He can be reached by e-mail at [email protected].