Next-generation Enterprise Information Systems
We can see a number of changes in market and society, enabled (some say ‘powered’) by advanced information and communication technology that strongly influence the way ‘the world turns’. These changes take place at a speed that cannot be compared to earlier changes in markets and society, i.e., they are more of a revolution than an evolution:
• stronger integration of both commercial and non-profit organizations to obtain economy of scale advantages.
• outsourcing of business processes to service providers (based on electronic market places)
• using information technology to increase efficiency of business processes, such as shifting front-end process completely to customer self-service and automatic retail of digital goods (most prominently music in digital format) without any human interference at the provider side.
• dynamically choosing partners for production processes or projects in complex supply chains (based on electronic market places) and integrating (automated) processes with them, such as temporary arrangements for providing adaptive, context aware end-user services.
• using real-time process information to make real-time operational (and tactical) decisions, such as monitoring physical goods flows (via RFID) for dynamic logistics management.
Broader scope of EIS
Traditionally, enterprise information systems were limited in their support for intra-organizational business processes. At the moment we see a growing importance of interorganizational (a.k.a. cross-organizational) applications of enterprise information systems in particular in e-business contexts. More and more, this includes dynamical aspects in external links, e.g. with respect to offered functionality and/or with respect to collaborating parties. Emergence of electronic markets for business functions facilities this form of dynamism. Service-oriented architecting is generally considered an important trend in this development.
To provide integrated solutions, new approaches must be developed to guide and manage the construction and maintenance of next-generation enterprise information systems (EIS). These next-generation systems are characterized by an overwhelming functional complexity combined with high requirements with respect to interoperability, flexibility, extensibility, adaptiveness and dependability. They will each employ many of the individual new technologies mentioned to cater for the complex requirements of the organizations they ‘serve’.
More and more we see that EIS are built of COTS-components (Commercial-offthe- shelf). In the past, these components were often integrated into one package by one vendor, but today we see that best-of-breed combinations of components of different vendors have a better fit to the business needs. To be able to develop, manage and maintain such systems, a clear and detailed architectural model of a system is required that is related to strategic objectives of an organization, or a network of collaborating organizations.
Architectural modeling has replaced classical specification and design, while programming has changed into configuration and orchestration. Components are replaced in a running system, so redevelopment is a continuous improvement process. Systems management is involved in sourcing questions: make, buy or rent components that deliver required services, as envisioned in business strategies.
As a consequence of the architecture development, the market place has changed dramatically: classical software houses became system integrators, programmers are replaced more and more by consultants that help their clients in selecting, configuring and integrating components. The vendor market is growing: software factories that build generic components for a specific functional area (horizontal market) or for a particular kind of business (vertical market). Both the integrators and the software factories have become much more professional during the last decade.
Research questions
Many research in computer science is demand-driven, i.e. inspired by problems that occur in the development or exploitation of software systems in practice. This is in particular the case in the EIS domain. We classify the research questions related to EIS according to the life cycle of an EIS.
• Business-IT alignment: How do we specify the functionality needed and how do we quantify the gap between the offered and required functionality? How do we verify if an information system satisfies conformance requirements from the business?
• Architecture: How do we describe an architecture of a system and what kinds of verification methods are applicable for architectures? How do we take into account that an architecture can be specified from multiple viewpoints?
• Componentization: How do we determine what useful components are and what the right level of parameterization is?
• Sourcing: How do we realize the components needed in the architecture? How can we share services? There are several sourcing approaches: build-time (software construction) and runtime (dynamic service selection).
• Configuration: What languages do we need for expressing parameters of components, what are the right methods for parameterization and what verification for parameterization is possible and useful?
• Integration: What is the right functionality of integration components, what languages do we need for integration, what are the best methods for integration are?
• Realization: What (combinations of) types of ICT are to be used for the realization of an EIS (embodiment of the architecture)? Service-oriented computing is an example of a recent and popular trend, typically using web services to achieve loosely-coupled system architectures.
Research
This section provides a short description of current research and possible new research directions.
• Business process management The main function of EIS is the support of business processes. Therefore the analysis, design and execution of business processes is a main topic of the EIS domain. The topic is called Business Process Management (BPM) or Business Process Integration (BPI). Note that the process models play a double role: at the business level in the enterprises as well as at the system level in the support by the EIS. Research focuses on development of models, architectures and automated support for dynamic interorganizational business processes (DIBP). During process enactment, collaborator organizations are found by searching business process market places and the sub processes are integrated with the running processes. Other topics of research are business process mining to discover process, control, data, organizational, and social structures from logged data; configurable process models as a more flexible method for reference modeling; analysis of process models for verification, validation and performance analysis (e.g., via simulation) using formal verification methods (e.g. Petri nets); adaptive process-aware information systems to enable dynamic changes of different process aspects and at different process levels. The goal is to maintain a close “fit” between real-word processes and the workflows as supported by IS, the current generation of which is known as Process-Aware Information Systems (PAIS); enterprise ontology based business process management as a fully implementation independent conceptual model of (the organization of) the enterprise that is comprehensive, coherent, consistent, and concise.
• Architecture and development The complexity of large and often inter-organizational information systems require a clear architecture, not only for the development of an EIS but in particular for management and maintenance of these systems. Architecture should be based on models: model-driven architecture (MDA). One of the main development problems is the integration of heterogeneous system components. A successful approach to architecture has it origin in web services: Service Oriented Architecture (SOA), where components are identified with the service they deliver. One of the main problems is the integration of heterogeneous system components. Typical research issues within this theme involve:
EIS architecture (re)design aims at developing knowledge on architecture principles, but also on the role of architecture in the software engineering process and in an organization. More and more, EIS are built from ready-made components that are bought or consumed as a service. Therefore, architecture design and implementation have to be integrated with sourcing processes. This requires deep technical insight in specification and qualification of architecture component.
Service-oriented integration aims at reuse and integration of existing system services, rather than building new systems from scratch. New services can be composed by reusing existing ones provided by various service providers. Today powerful integration tools (e.g. different kinds of message-oriented middleware) are available to overcome technical and syntactical heterogeneity. Yet, semantic heterogeneity remains as a major barrier to seamless integration of autonomously developed software components.
Enterprise architecture based development aims at developing a prescriptive notion of architecture which is complementary to the descriptive notion (blueprint, fundamental organization, etc.). There is a great need for developing formal, probably deontic logic based, specification languages for design, in order to be able to verify the consistency and coherence of a set of principles. Also desired is a compositional framework for architecture that represents the views and options of the multiple stakeholders. The framework should support what-if analysis to select a composition that fits the need of various stakeholders
• Maintenance and governance EIS have a long life cycle: 10 to 20 years is normal. The contribution of the development cost is in most cases less than 50% of the total cost of ownership. EIS will be modified during their life cycle so there will always be legacy parts. Reengineering will always play an important role in maintenance of EIS. The management of the evolution of EIS has become a serious task in which the alignment of the EIS to the business needs is an important aspect. research issues within this theme involve:
EIS maintenance and reengineering looks at methods to replace or modify parts of a system because it does not behave properly anymore, due to new user requirements or due to changes in the underlying platforms. The discovery of dynamical structure of software systems is one of the challenges. Different techniques can be used, among which process mining on executions of the software system is a new approach. Generation of parts of an EIS based on the (MDA) models obtained by reverse engineering of existing information systems is another new direction.
EIS governance addresses compliance of systems with legislation and security requirements. IT management processes require that a quantitative cost-benefit analysis is present for every EIS development or change project; methods for this are in their infancy. A future trend is to focus on performance (‘doing the right things’) in IT governance. This gives rise to quantitative models for analyzing application portfolios.
Governance of services in multi-actor networks addresses problems in the way dependencies among public agencies and shared service centre can be coordinated. Governance mechanisms address often an unbalance in value networks. Current research focuses on variety of governance mechanisms and principles to ensure that organizational networks behave in the right directions. The governance of shared services is still an unexplored domain and there are hardly any generic principles and mechanisms. More research is necessary into the types of governance models and which type of governance result in better performance and how networks can be governed.
