ICWS05 & SCC05 Panel 1: Services Science: Services Innovation Research & Education [Slides]
ICWS05 & SCC05 Panel 1: Services Science: Services Innovation Research & Education [Slides]
George W. Brown (Intel) [Slides]
Michael J. Carey (BEA) [Slides]
Tapas Kanungo (IBM) [Slides]
Akhil Kumar (Penn State University) [Slides]
Mohan Tanniru (University of Arizona) [Slides]
IBM Research, SERVICES SCIENCE: A NEW ACADEMIC DISCIPLINE? 2004.
Uday Kamarkar, Will you survive the services revolution, HBR, June 2004.
Paul Horn, The New Discipline of Services Science, BusinessWeek, January 2005.
James Tien and Daniel Berg, A CASE FOR SERVICE SYSTEMS ENGINEERING, JOURNAL OF SYSTEMS SCIENCE AND SYSTEMS ENGINEERING, Vol. 12, No.1, pp13. -38 M,arch, 2003
In a two-day session in May of 2004, starting with the theme of "The Architecture of On Demand Business", over 60 researchers from universities and IBM discussed a bold undertaking: develop and introduce a new academic discipline. A 120-page report was titled "SERVICES SCIENCE: A NEW ACADEMIC DISCIPLINE?", organized along a business model involving business strategy, people/workforce, business process, and underlying technology. The core message of the white paper is "Services has matured as a business as software once did, and there is a science underlying services that must be explored". The key challenge for academia and industry is to determine how to define and measure innovation. The government and companies need to invest in the research and development needed to "move services out of the realm of art and into the realm of science".
As an early mover in services science education, the School of Information Management and Systems at UC Berkeley is now offering a new course -- INFOSYS 290: Services Science and Business Engineering. As stated in the course description of INFOSYS 290, "The services sector dominates economic activity in most advanced industrial economies today. Similarly, "Web Services" and "Service Oriented Architectures" are important techniques and concepts as services are increasingly delivered over the Internet. But our scientific understanding of services is in a rather rudimentary state and the methods for designing and deploying them are ad hoc."
To help move this new field forward in services science research education, this panel assembles several experts in information technology and strategy in an open discussion that will attempt to provide new insights on the following potential questions (but not limited to):
What is services science, mangement and engineering (SSME)?
Our position is that business services can be represented at two levels which are different views of the same processes, one at the level of business semantics, the second as that of technical infrastructure. This two-tiered framework is referred to as the Integrated Process and Technology Framework. The first tier is comprised of business processes defined in terms of value chain reference models which can be used for process modeling, gap analysis, simulation, benchmarking and consensus building. The second tier is an architectural representation can be used that maps process models to components of the conceptual architecture and resources used for accurate, fast and flexible implementations of the services in a federation. The two independent but reconciled process representations facilitate the mapping of business process to core collaboration capabilities in a SOA.
Services account for over 70% of the GDP of most industrialized nations, over 60% IBM's revenue, and a tiny fraction of most academic curricula in the US. At IBM, by services we mean not only the computing infrastructure that perhaps provides services over the internet (e.g. on-line reservations, or retail), but also the human workforce involved in providing the services, the customer-facing employees, etc. So why study services? It turns out that in manufacturing of goods, one can exploit economies of scale and hence improve profit margins. However, in knowledge-intensive services, it is currently not possible to systematically improve profit margins in this way since a major part of the service involves interactions between human beings within the providers organization and across the client and provider organizations. Can we, then, understand the services client-provider "system" more systematically? Can we, for example use the tools that are used for understanding systems, and apply it to such socio-econo-computational systems and see where the inefficiencies are, predict the impact of possible policy changes, study the origins of variability, etc? At IBM, we believe that to address these issues we have to tackle the problem in a multidisciplinary way and have established a group of researchers with background in business, ethnography, and computer science and engineering to lead the effort.
While many aspects of services management have been explored, primarily in operations research (we know how to manage supermarket checkout counters, run hotels, and transportation services, etc.), yet services are assuming greater importance, and there is need to study them from new perspectives. A new development in the services area is the emergence of virtual organizations that add value by providing on-the-fly composite services. This panelist advocates a process-centric view of services. From a services point of view, it is necessary to not only model the sequencing and coordination of various tasks as current process modeling approaches do, but also include who is going to do the task, what is their availability, level of experience, etc. along with other resources required to do the task, such as data, equipment, facilities, etc. Current process description languages such as BPEL also fall short in this respect. In addition, they also neglect issues related to roles and authorizations and their security implications. These should form an important part of the process description in a services context. Research is also required in the area of metrics for evaluating the service in terms of quality and other measures of interest to a business.
Service is a physical phenomenon that is defined when two agents or groups of agents are involved in an exchange of resources (products, information, etc.). In the case of service, an unbiased observation is made of the "effort" made by each agent to meet the needs of the other, and this manifests itself in measures such as "response to inquiry or an order" (pre and post-engagement), "adherence to an agreement or a contract (SLA)", etc. However, some of these measures are perceptual and/or cognitive in nature (e.g. acceptable levels of service and customer satisfaction) and volatile (change with time and use). Thus, any systematic experimentation is only feasible, if this cognitive process or perception of "service" can be modeled for observation, influence and change.
Ever since firms moved out of transaction processing to decision support, the MIS field has been trying to understand for over two decades the decision process of users and how it is impacted by their use of IT, and expert systems research for over a decade has been trying to structure the problem solving process of decision makers using various knowledge extraction techniques. These two disciplines can be valuable reference disciplines to better structure the concept of "service." In addition, marketing can also be a reference discipline, as it has been trying to influence an individual's "perception" of value using various theories of innovation adoption and diffusion, and these theories can also provide some clues on how best to "influence" service perception.
If one were to take a scientific approach to define service, then service embedded when two agents interact has to be modeled using various relevant service features and their impact on both tactical and perceptual targets. Tactical targets are those that are relatively insensitive to changes in the way IT is applied and used, or easily measurable (e.g. back-up of data or size of storage in a service level agreement), and perceptual targets are those that are sensitive to IT usage (e.g. satisfaction with the information provided, response time of the system). The impact of various changes in the way services are performed and supported using IT over time are observed and fed back to influence the service features, so the service, as a process, can continue to adapt to changes in the service targets.
The bottom line is - I am not sure service science is entirely a different discipline as much as a way of combining computer science, with the
J. Leon Zhao is Professor and Honeywell Fellow in MIS, University of Arizona. He has Ph.D. from Haas School of Business, UC Berkeley. He taught at HKUST and William & Mary and worked as Staff Scientist in LBNL and as Research Engineer in Honeywell. His research has appeared in over 80 conference and journal articles including such topics as web services and services computing. He co-chairs the 15th Workshop on Information Technology and Systems, 2005 with Services Computing as its theme and is co-editing the special issue "From Web Services to Services Computing" for Information Systems Frontier.
George W. Brown joined Intel in 1994 and is currently a senior program manager within the ISTG Research group. He focuses specifically on methods and tools to ensure Intel reaches its goals in supply chain management by identifying opportunities to apply information technology in innovative ways to solve business problems and improve Intel business processes. He is also the past chairman of the Supply Chain Council and has represented Intel in external research and benchmarking activities as past chair of the SCC Research Strategy Committee and current Board member of the newly formed Value Chain Group.
Michael J. Carey is 1983 Ph.D. graduate of UC Berkeley. He joined BEA Systems in 2001 and is currently Technical Director for the Liquid Data at BEA Systems, which just released a new product offering in the area of enterprise data services. His previous work includes a year and half at a Silicon Valley e-commerce startup (Propel Software), five years at the IBM Almaden Research Center working on DB2, data integration, and XML/database technologies, and twelve years as a Computer Sciences faculty at the University of Wisconsin-Madison. He is ACM Fellow and member of National Academy of Engineering.
Tapas Kanungo is with Almaden Services Research at IBM's Almaden Research Center in San Jose, CA, where innovation for IBM Global Services is the focus. Human sciences, On-Demand Innovation Services (ODIS), deep industry knowledge of future trends, and operations technology are areas of active exploration. Prior to joining IBM in 2001 Tapas was Co-Director of Language and Media Processing Lab at the University of Maryland, College Park, where he worked for 4 years. Tapas received his Ph.D. in Electrical Engineering from the University of Washington, Seattle, WA, in 1996. He has published over 75 articles in journals, book chapters, and conference proceedings and is a Senior Member of IEEE.
Akhil Kumar is a professor of information systems at the Smeal College of Business at Penn State University. He received his Ph.D. from Berkeley, and has previously been on the faculties at Cornell University and University of Colorado, and also spent a sabbatical year as a scientist at Bell Labs, Murray Hill, NJ. He has done pioneering work in XML based workflows. His research interests are in workflow systems, e-services, distributed information systems and intelligent systems.
Mohan Tanniru is Eller Professor and the MIS department head, Eller College of Management, University of Arizona. He received his Ph.D. in MIS from Northwestern University in 1978. Prior to 2003, he coordinated and/or directed over 200 IT projects with major companies such as GM, DaimlerChrysler, EDS, Lear, Comerica and Compuware at Oakland University. He was on the faculty of Syracuse University between 1982 and 1997. He has published over 75 articles in journals, books and conference proceedings. He has consulted with Proctor & Gamble, Carrier- UTC, Bristol Myers Squibb, Tata Consultancy Services of INDIA, and Tata Infotech.