Software – International Development
Software in the Global Village
Two software projects—one to streamline processes in hospitals, the other to transform cell phones into multimedia platforms—illustrate the complexities of global software development.
The Soarian hospital information system: Developers in both the U.S. and in India work on the project
Next year, your teenager may be able to call all his buddies and talk to them at once by pressing a single button on a cell phone. In a few months, a hospital near you may cut the time it takes to generate a diagnostic report from 48 hours to 15 minutes. Nurses may soon have as much as 50 % more time for their patients.
What these technologies have in common is that they are the progeny of two vast software development projects that have been structured to harvest the know-how of specialized groups in a worldwide organization. Soarian, a comprehensive healthcare information system for hospitals that has already entered service, is a software tour-de-force built on some 3,500 man-years of research and development in the United States, Sweden, India, and Germany.
IMS, on the other hand, which stands for IP—(Internet Protocol) Based Multimedia Subsystem, is a visionary software system developed in Germany, Austria, the Czech Republic, Slovenija, Croatia, India, France, Finland, Britain, Belgium and Greece that could soon transform every communication terminal into a multimedia platform.
The largest software development project ever conducted by Siemens Medical Solutions, Soarian captures and tracks a patient’s clinical and financial data from hospital admission to release while making it available throughout the enterprise to all authorized personnel. The very first Web-based enterprise health care information system available anywhere, Soarian has a potential market "in the range of four to five billion dollars," says Dr. Siegfried Bocionek, Chief Operating Officer, Siemens Health Services, and Group Vice President of the Soarian Enterprise business unit. Soarian is now being tested at some 20 healthcare centers in the U.S. and Germany.
Mobile Media Center. IMS, a project involving years of work by as many as 250 software developers, is now being tested by major communication service providers around the world and is slated for commercial introduction in 2005 when the first IMS/SIP-enabled cell phones hit the market. (SIP, or Session Initiation Protocol, is a de facto standard that defines how Internet communications are initiated and terminated.)
Basically a service control infrastructure for all forms of communication, "IMS will establish Internet technology in the wired and wireless network environment" says Dr. Edward Scheiterer, head of IMS business line management at Siemens Communication (Com). He adds that IMS will also introduce "the concept of a session broker to mix and manipulate all types of media." Adds Johannes Schinko, Vice President, Core Networks, at Program and System Engineering (PSE), a Siemens software house based in Vienna, Austria, "IMS will make mobile communication as multifaceted as natural communication."
IMS and Soarian are outstanding examples of how software mega projects have come to be organized and managed. "Dividing a project like IMS is always a challenge," says PSE’s Schinko. "Of course, it would be easier to have everyone at one location—and cheaper to have everyone in India or China. But at the end of the day, the division of work depends on system architecture and the level of experience that different sites bring to the table. Those are the determining factors." Schinko points out, for instance, that a key part of IMS—something called the Media Gateway Control Function—was conceptualized jointly by two Munich-based Siemens Groups, ICM and ICN. "But," he explains, "the know-how for this component was primarily available at a Siemens site in Greece, and it was therefore reasonable to develop much of the software there."
Another Siemens location, Romsey, England-based Roke Manor Research, was tapped for its specialized knowledge in addressing and compressing messages for the Session Initiation Protocol (SIP). Only one part of the giant project—development of functionalities specific to Symbian, an operating system for cell phones—was turned over to an outside company. "That work was performed by Digia of Finland," says Schinko. "They had specialized knowledge in that area. But," he emphasizes, "for strategic reasons, key technologies are never given to outside companies for development."
Unlike IMS, which benefited from many existing communication standards and protocols, thus simplifying development of its software applications, most of Soarian was developed from scratch. Manpower requirements ranged from approximately 900 Siemens developers in the U.S. to 350 people in India and 50 in Sweden. "We had experience in Bangalore with developing the syngo diagnostic imaging platform," says Bocionek, who headed that project as well. "And in some specialized areas where Siemens Medical Solutions did not have experience at the time, we tapped (what was then known as) Siemens Nixdorf in Sweden."
The IMS multimedia platform for cell phones was developed by software engineers at 13 locations in 11 countries in Europe and Asia
Technology and Psychology. With most of its people in Malvern, Pennsylvania, and Bangalore, India, Soarian development struggled with basic technical problems. "We were often confronted with breakdowns in line carrying capacity when we transmitted code overnight from India to the U.S.," recalls Bocionek. "And management reviews were hampered by time zone differences. One group would be worn out after a day’s work, while the other group would not yet be fully awake." Naturally, coordinating so many people in different cultures was also a challenge. "People need to take the time to understand their counterparts in another culture," says Bocionek. "That means that team leaders need to develop strong personal relationships and be good at imparting complex ideas to their people." He cautions that size itself can become an obstacle. "It’s crucial that the individual R&D groups don’t become too large. I would avoid having more than 150 people at one location because anything more than that requires an added layer of management, which slows things down."
For the IMS project’s Scheiterer, architecture, processes and project management are the cornerstones of successful software development projects. "However," he adds, "good communication and clear responsibilities for all participants are a precondition for all of these. At the end of the day, a shared vision and common goals are the key."
But the flip side of that is that without good architecture, even the best teams would fail. So what is good architecture? "In the context of an international development project," says Schinko, "it is architecture that, when possible, separates functional blocks, allowing them to be developed independently of one another. That helps to avoid misunderstandings, and sets the stage for good motivation." Adds Scheiterer, "Good architecture minimizes overlap while improving end-to-end functionality and quality."
However, no project, no matter who the architect is, is flawless. "The step from logical structure to technical solution can be a big one," says Bocionek, pointing out that overlap has a way of creeping in. "There are parts of almost any project that are not self-contained and need to be used in other places. It is the nature of the beast. After all, good architecture should describe the real world—but that is never perfect."
Arthur F. Pease