Seminars 2007

December 18, 2007
Amal Algammal, Cairo University
Modeling Web Services for Automatic Composition and Verification

There is a great interest paid to the web service paradigm during the last several years. One of the most important problems is to enable the automatic composition of web services to provide a new value-added service. The problem of automatic web service composition is very tightly related to the description of web services. The main models that address the problem of automatic composition are: OWL-S, Roman, Mealy/Conversation and Colombo Models. Each of these models provides a framework for characterizing and describing web services. They also provide different automatic composition algorithms. In our research, we focus on the modeling of the automatic composition problem. The most recent and richest model is the Colombo model, as it unifies all the previously mentioned models. The first problem we consider is that Colombo adapts complex and low-level formalisms which are difficult even for experienced web service developers. The first contribution of our work is that we have used Standard UML 2.0 to simulate every aspect of the Colombo model. We have implemented a prototype for the proposed UML profile on Visual Paradigm for UML (VP-UML) version 6 CASE tool. The second problem is that Colombo is still a theoretical, conceptual model, which means that it does not have an associative language. The second contribution of our work is the proposal of a set of related XML document types that can serve as a core for a Colombo language. We have also proposed the transformation rules between the proposed UML profile and the pr oposed Colombo XML document types. These transformation rules are also implemented in our prototype. Here we assume that the composition algorithm proposed by Colombo accepts and produces XML documents that conform to the DTDs of the proposed XML documents. The third problem we consider is that after the composite web service is produced, there should be some method to analyze and verify that the resulting composite service performs what is expected from it to do, and ensure some important properties (like the system will never reach a deadlock). The third contribution of our work is the proposal to map the resulting composite service to Process Algebra (PA) formalisms to be able to utilize the sophisticated automated verification tools associated with PA. Mainly we propose to utilize process equivalence and model checkers tools of PA.

December 17, 2007
Michele Mancioppi, Tilburg University
Workflow and Changes

Workflow management is growing ubiquitous in enterprise-class information systems, as well as in a multiplicity of other fields. Workflows are used to model processes involving multiple parties, while interacting with other processes, and modify data and representations of business objects. Some of the uses for workflows are, among the others: execute and automatize business processes and health-care practices, model conversations among agents, study properties of programs in software engineering and modify them in refactoring.
In the literature, a "Change Algebra" is a theoretical framework that describes which kind of changes can be applied to workflows, and which kind of validation is performed on them. The importance of a theory addressing changes in workflows is paramount. Long running processes are almost guaranteed to change over time (for instance, to address new compliance regulations). Changes are typically applied to the process definition (the structure of the process to be executed), so that future processes will accommodate the new requirements. Applying changes to a process definition generates a new process definition; though, it is often needed to preserve certain properties of a process definition across changes: for instance, that the way a process interacts with our processes is not affected by the change. Moreover, often changes applied to process definitions must be also applied to instances of the processes that are already running. While applying the change is supported by some of the state-of-the-art tools for Workflow Management, the comprehension of how to validate those changes is still very little.
In this seminar, the following will be presented: the entities to be describe in a workflow model, a classification of the types of change to be described in the Change Algebra, and the supported types of validation. The aim of the seminar is also to gather feedback concerning which features should a Change Algebra offer to accommodate uses of workflows in multiple disciplines (with a particular emphasys on business process modeling and Service Oriented Architectures.

December 03, 2007
Erwin Fielt, Telematica Instituut
Exchange Design for Electronic Intermediaries

Electronic commerce offers business-to-business electronic intermediaries new design options for supporting the exchange between customers and suppliers. However, will customers and suppliers use their services and be satisfied with them? A major problem for intermediaries is making the right exchange design choices. This is a complex undertaking because of the many design options on the one hand and the diverse, and sometimes conflicting, interests of stakeholders to be considered on the other. Our focus is upon the critical exchange design choices that require trade-offs between the interests of customers, suppliers and the intermediary. We will develop an exchange design model and patterns for business-to-business electronic intermediaries. The model and patterns are based upon the findings of four in-depth case studies and discussed in relation to theory. The model comprises four themes: the role of the intermediary, the linkages between intermediary and customers and suppliers, transparency for customers and suppliers, and the novelty of the intermediary. The combination of a model and patterns for intermediaries that draw attention to critical exchange design choices is new, and can support an intermediary in designing balanced exchanges with a positive net effect for all actors and without unexpected side-effects.
Erwin Fielt is currently working at the Telematica Instituut (www.telin.nl) in the Netherlands. The Telematica Instituut was founded in 1997 as a Technological Top Institute. It is an international collaboration of companies, social institutions and knowledge organizations bridging fundamental scientific knowledge and its application in the business community and society at large. Erwin Fielt received his PhD from the Delft University of Technology. His research interests are ICT-enabled innovation, electronic business, business networks, business models, and business processes. He works with companies such as ING, Philips, IBM, and Hunter Douglas and universities such as Delft University of Technology the University of Twente and the VU University Amsterdam in R&D collaborations.

November 27, 2007
Cate Heeney, The Ethox Centre, Oxford University
Reuse of Health Administrative Records: The Ethical and Practical Challenges

The paper looks at the issues for National Statistical Institutes (NSIs) around the reuse of public sector administrative records for statistical purposes. This topic is of ever growing relevance to the research community and is currently a topic of debate for health researchers in the UK in relation to access to National Health Service Data for example. However, the necessity of retaining identifiers and the ability to share information across organisational and departmental boundaries raise privacy concerns in particular. As practical steps are taken in the UK to put in place the technical and practical solutions for data sharing this paper considers them and the ethical matters that they raise. The aim is to explore these issues for NSIs in UK and the Netherlands in order to draw lessons from the two countries, which are applicable to the research community and the issue of secondary use of public sector data more generally.

November 20, 2007
Maria Valeria de Castro, University Rey Juan Carlos
A Service-Oriented Approach for the Development of Web Applications: From the Business Model to the Web Service Composition Model

Currently the service-oriented computing is one of the major research topics in the field of software development and has brought about an evolution in Information Systems (ISs) themselves, as well as in how they are developed. Nowadays, fully integrated enterprises are being replaced by business networks where each participant provides to the others with specialized services. Thus, many IS, mainly those for the Web, are created as a means to offer services over the Internet, usually involving simple or complex business processes in which several participants collaborate. Consequently, several languages for the execution of business processes have been created; they allow composing several Web services offering new and more complex services. Nevertheless, these languages are not appropriates to be used in the early stages of the development process of IS. So, it is useful to have special methodologies or techniques that allow systems analysts to understand services from a business point of view, as well as to facilitate the design and development of the Web services composition. In this talk we introduce a service-oriented approach for the development of Web applications, called SOD-M (Service-Oriented Development Method). SOD-M defines a model-driven process, based on the MDA proposal, which starts from high-level business environment modelling and enables us to obtain a service composition design. The service composition model obtained from this method facilitates the transformation to specific languages for business process execution, reducing the development efforts of service-oriented applications.
Short Bio:
María Valeria de Castro is associate professor in the Department of Computing Languages and Systems at the Rey Juan Carlos University sited in Madrid, Spain. She received her M.Sc. degree in Information Systems from the National Technological University, Argentine in 2003 and she got her Ph.D. in Computer Science by the Rey Juan Carlos University in 2007. Her research interests include Web Services Engineering, Service Oriented Computing, Web Engineering and Model Driven Engineering. She has co-authored several publications in national and international conferences and journals and she has participated in several research projects.

November 6, 2007
Salima Benbernou, University Claude Bernard Lyon 1, LIRIS (Lyon Research Center for Images and Intelligent Information Systems)
An Agreement Based Approach for Privacy Evolution in Web Services

Web services are among the applications that involve closely the customer's private information. In order to take into account the privacy concerns of the individuals, organizations (e.g Web services) provide privacy policies as promises describing how they will handle personal data of the individual. However, privacy policies do not convince potential individuals to disclose their personal data, do not guarantee the protection of personal information, and do not provide how to handle a possible evolution of the policies. In this talk, we introduce a framework based on an agreement as a solution to these problems. We propose a “privacy agreement model that spells out a set of requirements related to consumer's privacy rights in terms of how service provider must handle privacy information. We define two levels in the agreement (1) policy level (2) negotiation level. A formal privacy model is described in the policy level. The framework supports in the negotiation level of the agreement a lifecycle management which is an important deal of a dynamic environment that characterizes Web services. Hence, the privacy evolution is handled in this level. A negotiation protocol is proposed to enable ongoing privacy negotiation to be translated into a new privacy agreement.

Dr. ing. Salima Benbernou is currently working as an associate professor in Database, Knowledge Representation and Reasoning Group since 2002 at University of Lyon. She was working in 1999-2000 as a consultant at Cap Gemini Ernst & Young corporation. In 2004 and 2005 she was a research scientist at CSE school, University of New South Wales, Sydney, Australia. She is a member of many national and European projects. Her research includes web service specificationand verification security and privacy in web services, temporal reasoning in distributed databases and knowledge bases.

October 22, 2007
Wolfgang Prinz, Director and Head of CSCW Research Department at Fraunhofer FIT, Germany
Do we need CSCW 2.0 to overcome our current communication problems and can the Pragmatic Web provide a solution?

In the past research on collaboration technologies has focussed primarily on the support of specific collaboration tasks and patterns. However, cooperative work is not organised in application specific tasks but processes and activities. Therefore approaches are required that integrate different applications in a cooperation context. In my talk I'll address current problems of the first generation of groupware and I'll outline first approaches towards a better integration from a system and user perspective. This may lead to a new generation of groupware following the vision of the Pragmatic Web to augment human collaboration effectively.

September 25, 2007
Sudha Ram, Eller College of Management, University of Arizona
Investigating Data Provenance in the Context of New Product Design and Development

Our objective in this research project is to define the semantics of Data Provenance. Data Provenance refers to the lineage of data in terms of various key events that occur during its lifecycle and other related information associated with the events. Provenance enables users to share, discover, and reuse data, thus streamlining collaborative activities and reducing the possibility of repeating dead ends. We have developed an ontological model of provenance called the W7 model that captures the semantics of data provenance as a combination of seven interconnected elements including “what”, “where”, “when”, “how”, “who”, “which”, and “why”. The element “what” describes events that affect data, including its creation, transformation, and archiving. The other elements are connected to “what” and describe different aspects of the events. The element “when” records the event time and “where” the event location. “How” documents actions leading up to the events. “Who” refers to people or organizations involved in the events. “Which” describes the instruments or software applications used in the events. Finally, “why” is defined as the decision rationale behind the creation, transformation, or other events that affect data. We have implemented a prototype software system to automate the harvesting and use of provenance. We are also applying the W7 model to various application domains including Product Life Cycle Management, and the Wikipedia to detect vandalism and to gauge data quality.

Sudha Ram is McClelland Professor, Management Information Systems in the Eller College of Management at the University of Arizona . She received a B.S. degree in mathematics, physics and chemistry from the University of Madras in 1979, MBA from the Indian Institute of Management, Calcutta in 1981 and a Ph.D. from the University of Illinois at Urbana-Champaign , in 1985. Dr. Ram's research deals with issues related to Enterprise Data Management. She has published articles in such journals as Communications of the ACM, IEEE Transactions on Knowledge and Data Engineering, Information Systems, Information Systems Research, Management Science, and MIS Quarterly.

Her research has been funded by organizations such as IBM , Intel Corporation, Raytheon, US ARMY, NIST, NSF, NASA, SAP, and Office of Research and Development of the CIA. She serves on the editorial boards of several journals and has chaired several workshops and conferences supported by ACM, IEEE, and AIS. She is a cofounder of the Workshop on Information Technology and Systems (WITS) and serves on the steering committee of many workshops and conferences including the Entity Relationship Conference (ER). Dr. Ram is a member of ACM, IEEE Computer Society, and Association for Information Systems (AIS). She is also the director of the Advanced Database Research Group based at the University of Arizona. Her most recent honors and awards include an IBM Faculty Development Award and UA Leading Edge Innovator in Research Award in 2007. More information about her provenance work you will find at http://kartik.eller.arizona.edu

September 11, 2007
Jian Yang, Macquarie University, Sydney
Supporting Differentiated Services With Configurable Business Processes.

In order to provide flexible and reusable business services, it is desirable to offer users or applications the same service with different service quality, different interaction paths, or different outcomes. We call this design principle as Service Differentiation. In this talk we present a working service design method where variability is externalized as business policies so that the business process(es) does not need to be altered for any anticipated changes. Service differentiation is realized by configured business processes and interfaces, and the dynamic 'binding' between user/application with a specific interface is determined by business policies during service invocation time.

June 19, 2007
Jose J. Gonzales, Agder University College, Norway
Long Term Instability in Constituency and High Priority Workload in Computer Security Information Response Teams

Persistent, long-term instabilities in the high-priority incident workload of computer security incident response teams (CSIRTs) cause serious problems. When in overshoot mode, the available resources are overstretched and the quality of service suffers; when in undershoot mode, there is waste of available resources. Using data provided by one of the oldest coordinating CSIRTs we developed a system dynamics model that reproduces the behaviour over time, provides a causal explanation for the phenomenon and allows analysis of CSIRT management policies. Three main feedback loops are responsible of an S-shaped behaviour followed by oscillating overshoot and undershoot. Growth in the workload is driven by a reinforcing word-of-mouth process; a delayed balancing loop of quality of service counteracts growth in the workload; another balancing loop, driven by the constituency turnover, gradually restores quality of service. Perception delays are crucial for the dynamic behaviour. An improved communication of the service level provided by the CSIRT is the most effective policy to mitigate long-term instability. The results are arguably generic and might be applicable for other coordinating CSIRTs.

June 08, 2007
Kishore Sengupta, INSEAD
What isIT Agility? Does it affect Organizational Performance? A Conceptualization and Some Empirical Evidence

IT agility – the ability of a firm to adapt its IT capabilities to market changes - is increasingly suggested as an important organizational capability. Yet, a review of the relevant literature suggests that the construct is still ill-specified, that it lacks reliable measurement and that its posited contribution to organizational performance needs empirical validation. This paper aims to extend IS research by providing insights on IT agility and its relationship with organizational performance. Based on a synthesis of research in IS and related fields, we posit that IT agility is of two different types: range- and time-agility. We utilize this conceptualization to propose an approach for measuring IT agility in organizations, and empirically assess the impact of different types of IT agility on organizational performance in large manufacturing companies. Our findings show that both range- and time-agility have a positive relationship with organizational performance. However, there are two sets of bounds. First, organizations that exhibit high range- and time-agility have difficulty in extracting value from the combination. Second, the relationship between time-agility and performance is inverse U-shaped. That is, increasing time agility beyond a point can be counterproductive to performance. We also find that the impact of IT agility on performance is moderated by two variables – the ease with which a particular type of agility can be attained, and the dynamism of the environment. The relationship between range-agility and performance is positively moderated by ease, i.e., range-agility creates more value for an organization when it is easier to attain. In contrast, ease negatively moderates the relationship between time-agility and performance. That is, companies are liable to over-speed when time-agility is easy to attain. Finally, market dynamism positively moderates the link between time-agility and performance, thereby demonstrating that time-agility is more valuable in highly dynamic and unpredictable environments. In contrast, the moderating effect on the range-agility performance link is negative, thus showing that range-agility ismore valuable in less dynamic and predictable environments. Thus, when it comes to IT agility, more is not always better. Rather, the benefits of range- and time-agility are bounded, and are contingent on ease and market dynamism. The results of the study provide guidelines for appropriate fits in the configuration of: the type of IT agility, the ease with which it can be attained, and the organization’s external environment; thus enabling managers to make decisions on investment and capability enhancement for IT agility. We also propose several avenues for future research.

May 15, 2007
Olaf Zimmermann, IBM Zurich Research Laboratory
Reusable Architectural Decision Models asMicro-Methodology for Service-Oriented Analysis and Design

During the construction of service-oriented architectures, service modelers concern themselves with the characteristics of good services and how such services can be designed. For instance, they look for advice regarding interface granularity and criteria to assess whether existing software assets are fit for reuse in service-oriented environments. There are no straightforward answers to such questions – service identification, specification and realization techniques are required. Service identification and specification are well covered by existing methodologies; for service realization, architectural decision models can be leveraged. In practice, the construction of architectural decision models is an education- and labor-intensive undertaking; if such models exist at all, they often are isolated from other artifacts. In this talk, we present a new engineering approach to service modeling that leverages reusable architectural decision models as its central service realization concept. The approach is based on a multi­level architectural decision model harvested from industry projects. The benefits of service engineering with reusable architectural decision models are semi-automatic decision identification in analysis models, higher decision making quality, and improved decision enforcement and risk mitigation capabilities. For illustration purposes, we present two case studies, the BPEL enablement of the order management application of a telecommunications wholesaler, and the Web services externalization approach followed by a shared service provider in the finance industry. We will look at drivers for the projects as well as the solution architectures, and share some of the lessons learned.

February 27,2007
Michael Weiss, Carleton University
Business Processes, Aspects, and Patterns using URN

The User Requirements Notation (URN) combines a visual scenario language (Use Case Maps) with a language for goal modeling (Goal- oriented Requirement Language). Its purpose is to model both functional and non-functional aspects of a system at the requirements level. Recently, we extended the notation to include the concept of aspects. Aspects improve the modularity, compositionality, reusability, scalability and maintainability of URN models. In this talk I will present our initial work on creating URN models that represent both business processes and associated non-functional concerns (such as security) using the proposed aspect-oriented extensions. Recurring supporting business processes can be documented in the form of patterns. Patterns can also be adapted from existing patterns for object-oriented systems. I will also touch on our work on business process monitoring using URN. The results reported were obtained in collaboration with Gunter Mussbacher and Daniel Amyot from the University of Ottawa, and several master's students.
Bio:
Michael Weiss, Associate Professor of Computer Science and a member of the Technology Innovation Management program at Carleton University in Ottawa.

February 01, 2007
Stijn Oomes, Delft University of Technology
Making friends during disaster response

Our goal is to improve the performance of teams and organizations by designingand developing collaboration support systems for them. One of our main interests is disaster response, especially the challenge of coordinating the joint activity of alarge number of people in such an ad-hoc and diverse organization. We want to support what we have called “organization awareness”, or the understanding of the people and parties that make up the organization, and how they relate to each other in terms of roles, responsibilities and tasks. We want to provide them with visualizations, analogous to geographical maps, that go beyond thestandard organizational charts. Since we take a human-centered approach, I will first discuss the sociology and psychology of collaboration and coordination, before switching to design mode.

January 18, 2007
Paul Johannesson
Royal Institute of Technology in Stockholm
From Business to Process Models - a Chaining Methodology

In this seminar we discuss the problem of how to go from a business model to a process model in a systematic way. Business models are economic models used for business analysis, while process models capture low-level business activities and their coordination. We propose a method that starts with a business model where the main actors and their relationships are identified. This forms a basis for design of a final process model. Processes are described in terms of patterns stored in a pattern library.

See also

• Previous Seminars 2006