A complex adaptive_system_perspective_of_enterprise_architecture_in_electronic_government


Published on

Published in: Technology, Business
  • Be the first to comment

  • Be the first to like this

No Downloads
Total views
On SlideShare
From Embeds
Number of Embeds
Embeds 0
No embeds

No notes for slide

A complex adaptive_system_perspective_of_enterprise_architecture_in_electronic_government

  1. 1. Proceedings of the 39th Hawaii International Conference on System Sciences - 2006 A Complex Adaptive System Perspective of Enterprise Architecture in Electronic Government Marijn Janssen George Kuk School of Technology, Policy and Nottingham University Business School, Management Jubilee Campus, Wollaton Road, Delft University of Technology Nottingham, NG8 1BB, Jaffalaan 5, NL-2600 GA, Delft The Netherlands United Kingdom Tel. +31 (15) 278 1140 Tel. +44(115)8466611 Marijnj@tbm.tudelft.nl g.kuk@nottingham.ac.uk Abstract related development work that can be undertaken for enterprise integration [23]. For example, at the local Within the public sector, the deployment of level, governmental agencies have the autonomy toenterprise architecture is often an attempt to address design their own information architecture so long asthe decentralization/centralization relationships to they adhere to the existing laws and financialimprove the links between the central and the local regulations. This autonomy can lead to massive localgovernments. The underlying aim is to provide a better differences as the centrally advocated architecturalstructure to manage the diverse, independent and localIT-related projects and development activities. design principles and standards may not be adopted In this paper we propose a novel approach to [16].analyzing and understanding the requirements and Due to the local diversity in terms of IT-basedlimitations for enterprise architectures in government. architectures and the number of vendors involved inWe use the perspective of a complex adaptive system the design and development of IT-based systems,as a metaphor to examine 11 e-government projects increasingly central administrators and managers arewith each involved the development of enterprise finding it difficult to manage and effect the neededarchitecture in the Netherlands (1980s - 2004). transformation and modernization [33]. Specifically,Through analyzing the key interaction points between they are struggling in dealing with the complexity andthe central and the local governments, we identify problems associated with independently developed andarchitectural design principles that will increase inter- managed projects at the local level. As localorganizational jointness and IT implementation governmental agencies have a large degree ofsuccess. autonomy, they are often outside the central purview and control. Inevitably, systems developed and1. Introduction decisions taken in local governments are often out of sync with the central government. Also the ways in which the central government interacts with the central It is often stated that the implementation of governments can increase the final complexity thatenterprise architecture offers a way forward in enterprise architecture has to deal with. Simon [31] (p.integrating independent IT silos across inter- 486) asserted that a complex system is one “made up aorganizational agencies (e.g. [7],[20]). Within the large number of parts that interact in a nonsimplepublic sector, the joining-up remains at the top of e- way.” Hence, a key objective for using architecturalgovernment agenda of most of the Western countries. concepts is to manage the complexity and contain theAs part of the modernization plan, governments seek increasing rate of change [5].to offer their citizens a seamless service delivery by As targets are increasingly difficult to formulate andlinking various independently developed systems and the path to a long-term strategy can change as theapplications. Yet the development and implementation target shifts [1], enterprise architecture needs to adaptof enterprise-wide technologies is a massive and to the change-on-change element that characterizes thecomplex undertaking [17]. surrounding environments. The large number of actors One of the key obstacles is the legal issues may involved both at the central and local levels, and theconstrain the extent of business transformation and IT- dynamic and nonsimple interactions can hamper the 0-7695-2507-5/06/$20.00 (C) 2006 IEEE 1
  2. 2. Proceedings of the 39th Hawaii International Conference on System Sciences - 2006decentralization and improvement of links between the 2. Theoretical Backgroundcentral and local government. In this regard, publicorganizations not only have to stay adaptive to the 2.1. Enterprise architectureconstant change but also manage complexity attributedto the interactions between the central and the local Enterprise architecture lacks a universally acceptedgovernments. definition [30]. It is a kind of city plans that detail To understand the underlying interactions, we use policies and standards for the design of infrastructurethe perspective of a complex adaptive system (CAS) as technologies, databases, and applications [14].our empirical lens to examine the interactions between Architecture aims at creating some kind of structure inthe central and the local governments. The essence of a chaotic environment using systematic approachesCAS is the study of systems built of individual agent [1]. The term “enterprise” refers to the scope of thethat are capable of adapting as they interact with each architecture, dealing with the organization as a wholeother and with an environment, and especially the or multiple agencies rather than with a certainattempt to understand how the individual affect the organizational part or individual components and/orsystem-level responses [2]. Most work in CAS has projects [8]. The architecture is based on statements ofbeen conduct in highly abstract and artificial systems, how an enterprise wants to use IT, not on what andlike cellular automata and genetic computer how information has to be made available. Thealgorithms, and in the fields of economics, sociology, strategic aspects of IT systems provide the contexts formicrobiology and medicine [26]. In recent years, CAS the architectural design choices and decisions.has attracted much interest in management and Enterprise architecture models provide ways to dealorganizational related literature (e.g. [9],[10]). It has with the complexity including work (who, where),been successfully applied to explain the adaptive function (how), information (what) and infrastructurebehavior of biological systems, optimization tasks and (how to) [5]. The enterprise framework formula, ininductive learning [10]. general terms, specifies how information technology is We use the general definition of CAS as a system related to the overall business processes and outcomesthat emerges over time into a coherent form, and of organizations, describing relationships amongadapts and organizes itself without any singular entity technical, organizational, and institutional componentsdeliberately managing or controlling it [18]. Through of the enterprise [12] [29].the CAS lens, we examine the types of interaction The idea of enterprise architecture is that it can bepoints afforded by the development and used to guide design decisions and limits the solutionimplementation of enterprise architecture. Our space by setting constraints. Architectural principlesintention is that by conceptualizing enterprise are textual statements that describe the constraintsarchitecture as CAS, public administrators and imposed upon the organization, and/or the decisionsmanagers can acquire a better understanding of taken in support of realizing the business strategies.interactions, and implement design principles that Principles restrict architectures and set the direction forincrease inter-organizational jointness and IT the future. Architectural descriptions can form theimplementation success. As such we use CAS as a basis for the implementation and transformation ofmetaphor to analyzing the key interaction points existing structure into the desired architecture.between the central and the local governments, and toidentify architectural design principles. This paper analyzes 11 e-government initiatives 2.2. Complex Adaptive System Theorywith each involved the development towards enterprise Desai [10] introduces the term adaptive complexarchitecture in the Netherlands between 1980s and enterprise (ACE) to differentiate from Holland’s [18]2004. The rest of the paper is structured as follows. In complex adaptive systems. This new term depicts thatsection 2, we review the literature about the many human-designed systems and processes are notrelationship between enterprise architecture and the complex and the focus should be on the complexity ofcomplex adaptive system theory. In section 3, we non-designed processes, interactions and relationshipsdiscuss our research approach. In section 4, we in addition to systems. They argue that a keyintroduce and describe the developments and key managerial task is to design adaptive systems orinteraction points within enterprise architecture for the processes that bring back or push forward conditionspublic administration in the Netherlands. In section 5, that are far from equilibrium into state of equilibrium.we derive architectural design principles and According to Lewin and Regine [22], the best-runguidelines based on the logic of CAS. In section 6, we organizations function much like a flock of birds, indiscuss our findings. Lastly, we discuss future research which individuals follow simple rules, interact withdirections in section seven. 2
  3. 3. Proceedings of the 39th Hawaii International Conference on System Sciences - 2006others to form a cohesive and dynamic whole. They aggregate behavior and helping the CAS to behave in afurther assert that complex adaptive systems emerge as predictable way [9].a result of the interactions among members. A CAS 10. Nested Systems: Most systems are nested withinhas the following emergent characteristics [13]: other systems and many systems are systems of1. Emergence: Agents in the system interact in smaller systems.apparently random ways. From all these interactions, If the organizations want to harness the benefits of apatterns emerge, and inform the behavior of the agents complex adaptive system, then the design of enterprisewithin the system and the behavior of the system itself. architecture has to balance between excessive and no2. Co-evolution: All systems exist within their own controls, and allow flexibility and adaptability suchenvironment, and they are also part of that that systems are not frozen because they are tightlyenvironment. As their environment changes, they need constrained or disintegrate due to little order [10].to change to ensure best fit. But because they are part Often the behavior of complex adaptive systems can beof their environment, when they change, they change predicted using a few simple rules [9]. In this respect,their environment, and so it goes on as a constant using simple rules, enterprise architectures shouldprocess. constrain the actions that can be taken and give3. Sub optimal: A complex adaptive system does not guidance for taking the right direction, in such a wayhave to be perfect in order for it to thrive within its that it is still possible to adopt and evolve as theenvironment. Once a complex adaptive system reaches environment changes.the state of being good enough, it will trade offefficiency with greater effectiveness. 3. Research approach4. Requisite Variety: The greater the variety withinthe system, the stronger it is. Complex adapt systems To explore the nature and background of enterpriserely on ambiguity, paradox and contradictions to create architecture in the public administration of thenew possibilities. Netherlands, we conducted comparative case studies.5. Connectivity: The ways in which the agents in a We gathered data primarily through semi-structuredsystem connect and relate to one another is critical to interviews. Following the recommendation ofthe survival of the system, because it is from these Eisenhardt [11], we used a theoretical samplingconnections that the patterns are formed and the approach to guide our case selection and to refine ourfeedback disseminated. The relationships among the research questions. With an emphasis on cross-caseagents are generally more important than the agents diversity, we selected 11 e-government projects withthemselves. each involved in the development of enterprise6. Simple Rules: Complex adaptive systems are not architecture.complicated. The emerging patterns may have a rich Each case involved a retrospective, in-depthvariety, but like a kaleidoscope the rules governing the examination of at least 2 key informants’ account offunction of the system are quite simple. the contexts and rationales of how the enterprise7. Iteration: Small changes in the initial conditions architecture was developed. Many interviews wereof the system can have significant impacts after they directly involved or experienced the results of multiplehave passed through the emergence of feedback loops. projects. In total, we interviewed 25 informants, and8. Self-organizing: There is no hierarchy of each interview lasted between 1 and 3.5 hours.command and control in a complex adaptive system. Especially the interviews involving multiple projectsThere is no planning or managing, but there is a lasted for more than one and a half hour. The interviewconstant re-organizing to find the best fit with the transcripts were then analyzed to identify the contextsenvironment. The ability of the agencies to intervene underpinning the key interaction and contact pointsin the course of events is a key characteristic of CAS between the central and the local governments, the key[9]. actors involved in the development, and the business9. Edge of Chaos: Complexity theory is not the same and technical logics behind the design choices andas chaos theory, which is derived from mathematics. solutions. To validate any claims and factualBut chaos does have a place in complexity theory in information, we also collected data from multiplethat systems exist on a spectrum ranging from sources including follow-up phone interviews, internalequilibrium to chaos. The most productive state to be memos, website research, publicly availablein is at the edge of chaos where there is maximum information and other relevant consultancy reports.variety and creativity, leading to new possibilities.Controls, such as rules and regulations or institutionaland budgetary restrictions, ensure that an agent’sbehaviors is limited, in this way changing the 3
  4. 4. Proceedings of the 39th Hawaii International Conference on System Sciences - 20064. Case study: development of an place to monitor and govern the development at theenterprise architecture municipality level. Although the central government tries to guide this Figure 1 displays the networks of agencies: the process by setting national targets, the top-downnational agencies at the top and the municipalities at approach is often criticized for not taking the localthe bottom. The focus of the present study was on the realities into consideration; and the unrealistickey interactions between the Ministry of Interior and timescales for the realizing those targets often result inthe municipalities. a rushing mentality, which in turn has led to poor quality and substandard implementation. With the advent of e-government, the Ministry of Interior has National level decided that they want to have more grip on the system architecture of local municipalities. Against this backdrop, the Ministry has started since 1980s toother develop enterprise architecture as a means of directinggovernmental constituents and constraining the architectures among the Dutchagencies municipalities. 4.2. Overview of developments Starting from the eighties, the emphasis was on automating the citizens’ records by enabling local municipalities to process and authenticate information level in relation to applications for passports and social security. The most important initiative in this regardFigure 1. The network of agencies was the creation of standards for the GBA (in Dutch: Gemeentelijk Basis Administratie). The GBA is the4.1. Background Dutch vital record registration of data about all residents living in a particular city or village. Within the existing laws and financial regulations Also due to mergers and consolidation of theof the Netherlands, the local governments enjoy a high software market, two software vendors became thedegree of autonomy in making their own decisions. dominant suppliers and had effectively created aThey have their own budgetary control and can make duopoly for the municipality market. The technicalindependent IT investment decisions. The local dependency and the contractual lock-in had madegovernments are hierarchically organized with varying switching to another vendor very costly. One of thedegrees of coupling and autonomous coordination. architectural aims was to reduce this duopoly by Despite all the good intentions and efforts, the opening up the technical lock-in through the use ofexisting efforts to direct IT investment and modular components and standardizing interfacesdevelopments are often ineffective. This has resulted in among the components.a highly fragmented and heterogeneous architecture In the mid-nineties, public managers became awarewithin the organization with disparate business of the need for directing developments towards e-processes, applications and operating systems. The government. This has resulted in a number of planssystems in place have different capabilities, data and revised plans of the ministry of interior andformats, content, and user-interfaces. There is no kingdom relationships (e.g. [5][6][15][25]). Table 1uniformity and interoperability is limited. provides an overview of the projects initiated at the System development projects can be characterized national level and the municipalities’ responses. Thisby the silos-like behavior including redundancy and table is limited to the main projects at the nationalduplication of efforts, and the lack of reuse and level, as the number of projects is far greaterinformation sharing. The silos can be attributed to the considering that the number of local agencies is almostautonomy that individual municipalities have in 500 and the number of municipality departments isexaggerating their unique business requirements and even higher.therefore demanding for unique systems. And thesystems got out of hand as there exists no structure in 4
  5. 5. Proceedings of the 39th Hawaii International Conference on System Sciences - 2006 Table 1. Overview of national initiatives and local-level responsesYear National level initiated project Local government response IT implementation Outcome/Performance1980s Standardization by founding a Municipalities developed and started operating Upside: Clear data ownership citizenry authentication registry. the citizenry registry. Overtime the resulting in higher data This registry defined the data to be development and maintenance was shifted to quality, improved stored in the registry but not the commercial software vendors and all kinds of interoperability and additional system modules and functionalities. extensions were developed to provide services functionalities. based on data. The software market for municipalities is characterized by the duopoly Downside: An increase in the lock-in. level of technical and strategic lock-in, and the number of systems in place.1995- OL2000: development of a Local customization and deviation among Upside: Standardized products2000 standardized product and service agencies of the use standardized product and and services catalogues. catalogue to enable the creation of service catalogues. Some organizations one-stop shop by municipalities extended this catalogue, as it did not contain Downside: Customization policy-making or law-enforcement related deviated from the standards. product and services1998 The parliament set the target that Local governments start developing websites Upside: Clear targets. 30% of the content and services on a one-to-one basis. A number of software should be available online by 2001. vendors developed web information systems Downside: Duplication of for use by government agencies. The number development projects; and variety of systems increased to meet this creating of interoperability target. problems, no fostering of reuse; and no sharing of experience2000- Re-development of business registry Municipalities were involved in the Upside: Clear data ownership that previously operated by over development of the business registry, yet resulting in higher data quality 500 municipalities. The aim was to currently there is still no consensus how the and additional functionalities. connect all the municipalities by interface should look like. The tendency seems 2004 and to enable to access all to be to define the building blocks and Downside: An increase in basic information at a national level. interfaces, and let multiple software vendors lock-in and the number of build a registry. systems in place.2001- Selecting and support of ‘super The evaluation of these pilots was very Upside: Fostering local pilots’. A limited number of large disastrous. The developed components were development. municipalities were given a large not suitable to reuse, the reports describing the amount of money to develop certain experiences remained too vague and the Downside: limited reuse for components of the architecture (e.g. “super pilots” were lagging behind on other other municipalities, not authentication, back-office municipalities who did not get any funding. dealing with interoperability integration, etc.) problem.2004- The parliament set the target that As with the target set discussed in the row Upside: Clear targets. 55% of the transaction service above, this resulted in an increase in the should be available in 2005, and variety of systems in place. Downside: Duplication of 65% of the services should be development projects, available online by 2007. creating of interoperability problems, and no reuse and/or learning transfer.2002- Founding the ICTU, an over- Municipalities find it difficult to adopt the Upside: Clear focal point for coupling architecture program reports, systems and concepts provided by the e-government enterprise office running all the nation-wide ICTU. They have a lack of knowledge to architectures. programs including: open source interpret the reports and translate them to their and standards, and infrastructure own situation. Downside: For municipalities component developments. it is not clear how the results and reports of the ICTU can be used2004 Data cannot be easily exchanged as Municipalities will be forced to obtain Improved interoperability, various definitions are used and business data from the business registry, clear data ownership resulting could be of low quality. Founding which will be operated by the chambers of in higher data quality. 5
  6. 6. Proceedings of the 39th Hawaii International Conference on System Sciences - 2006 of six authentic registries for commerce. Businesses provide changes in data citizens, business, addresses, only to the chamber of commerce. This is Downside: no additional building, land and maps. The similar like citizens only have to provide functionalities and interfaces business registry should be adopted changes to the municipality they are living in, are included which might in 2007. operating the citizen’s registry. result in a large variety of All government organization must systems having different reuse the information stored in the granularity six authentic registries by 2007. The control of the enterprise architecture was 4) Authentic registration: the Founding of sixtraditionally scattered around national agencies. One authentic registries for citizens, businesses,interviewee remarked that “we have not only one buildings, land, addresses and maps.control body; we have at least 10”. Consequently, the 5) E-form generator: defining a generic formMinistry of Interior sought to create a Program Office generator for generating standard as well asthat would oversee, approve, and coordinate the many customized formsindependent and uncoordinated efforts of the ISdepartment, and hence the founding of the ICTU. Its Besides the development of these basic buildingfirst mission was to concentrate all development blocks there are several other programs to helpactivities and other related efforts under one program municipalities (i.e.office. The next step was to ensure that there is a http://www.ictu.nl/activiteiten.html), such as opensystematic and coherent relationship among the source and standard (OSS) program and transactionvarious programs. This was done by defining basic port. OSS is aimed at stimulating the use of openinfrastructure building blocks, which included: source software by providing knowledge about open1) DigiD: A national authentication service for source concepts and the stimulation of the citizens and business combining on a number of development of standards. The transaction port is existing authentication projects initiated by functioning as a channel for exchange information various agencies. between businesses and governmental agencies.2) Public key infrastructure: creating an Over time it was recognized that it was important infrastructure for facilitating transaction between that initiatives should be started at the local level. governmental agencies. Table 2 shows the initiatives emerging at the local3) Citizens service number: Defining a unique level and which were adopted at the national level. The number for each citizen to gain access to it’s uptake of emerging initiatives is inline with CAS personal data by all governmental agencies. theory. Table 2. Overview of local level initiatives and responses at national levelYear Local level initiatives National level response IT implementation Outcome/Performance2001 One municipality developed a web- Founding the EGEM by the information Upside: Close to intake system what was regarded as managers of the Dutch municipalities to organizational realities at the a best practice. To recoup the collect and disseminate best practices. The local level. investment costs, they started EGEM was a project within the ICTU. The marketing and selling their system EGEM was assimilated by the ICTU and is Downside: Agencies can still to other agencies. Based on this idea now a separate program. select other systems. EGEM was founded.2000 In the early years of this century, The national level reacted to the emerging of Upside: Concentration of some municipalities started merging ICT-departments and sharing of services at a efforts, lowering costs, access their ICT-departments. They found local level, by looking at opportunities to to more expertise. their ICT-budget too limited to meet develop infrastructural building blocks that the targets set by the national can be shared by local agencies. Downside: Remaining at a government. local level and sharing is limited2002 Each municipality had to develop Within the ICTU a new program was started Upside: Reuse of their own authentication service to to build an authentication service. Bundling of functionality, concentrating provide access for citizens. This various authentication efforts into a basic development. could range from a login using name infrastructure building block for creating a and zipcode to the use of identity nation-wide authentication service. cards. 6
  7. 7. Proceedings of the 39th Hawaii International Conference on System Sciences - 2006 Current developments seem to have the tendency to breed variety, however, might not result in the desiredcentralize critical activities after they emerged at the outcomes as shown by the fifth super-pilots projectlocal level, but nevertheless still giving enough Budget was allocated, however, the conditions tofreedom to the decentralized level. Which is spend it were not provided. The results obtained fromconforming the ideas of CAS theory. For this purpose the super-pilots projects were not reusable by otherinteraction points between agencies need to be agencies.fostered, but interactions should also be directed by Variety is important, but should be created withconcentrating them. Without concentration the care. The conditions, such as reusability and theinitiatives might not be aware of each other and create justification for spending the money, should be definedmuch more diversity than is needed and manageable. without disputing the autonomy of the initiatives.We also found that breeding diversity is important asthis results in better outcomes, however, too much 2) Set no targets without providing constraintsdiversity results in large interoperability problems. The targets set by the parliament and the Ministry of Interior strongly encouraged the municipalities to5. Architecture design principles adopt. They developed systems, made their content available online and started developing transactional In a CAS, agents interact in an unpredictable and services. As such, the informants found it as a veryunplanned way. CAS predicts that the agent can never powerful mean to direct initiatives. The setting ofoptimize as the space of possibilities is too vast and targets without providing any constraint resulted in athere exists no practical way of finding the optimum. variety of heterogeneous systems and accompanyingHowever, from those interactions, regularities emerged interoperability problems. From an architectural pointand form a pattern [18]. Choi et al. [9] state that of view, this can be considered as undesirable. Eachcollective influence of schema in a CAS follow a city council allocated different amounts of budgets topareto principle, i.e. a few dominant principles dictate its ICT-department to accomplish this target, however,the vast majority of behavior. Thus, the generative all at the same time were duplicating efforts. Eachmechanism organizing a CAS is often referred to as “a department had to reinvent the wheel and there was nofew simple rules.” coordination among efforts. A more fruitful direction Hereafter, we derive some simple rules grounded in seems to be to constrain the number of directions thatCAS theory, based on the analyses of the 11 projects. can be taken.We label them as architectural design principles orguidelines, as they are textual statements that describe 5.2. Concentration of activitiesthe constraints imposed upon the organization, and/orthe decisions taken in support of realizing the business A key characteristic of CAS is that there is nostrategies and can be used to guide future projects. The hierarchy of command and control in a complexprinciples are based on the primarily findings that adaptive system. The best initiatives might besolely breeding of diversity is not sufficient, there is a disseminated better by concentrating efforts andneed to have a focal point to concentrate activities and creating awareness of this concentration point. This isefforts and an increase in the number of interactions clearly support by ICTU, which concentrates all e-result in new and better initiatives. government related programs within one organizational entity.5.1. Solely breeding diversity is not sufficient 3) Stimulate growth of successful projects CAS states that the greater the variety within the The basic idea behind this principle is to breedsystem the stronger it is to use contradictions to create initiatives that might become successful and result innew possibilities to co-evolve with their environment. best practices. This also means that projects targeting similar areas and not likely resulting in success should1) Stimulate diversity/ breed diversity not be supported and discouraged. The current policy The first six projects show that solely breeding is to have several overlapping projects. By using adiversity is not sufficient for an effective governance successful municipality as a reference point for otherof enterprise architecture. The stimulating of diversity municipalities, the resources can be concentrated andresulted in a large number of heterogeneous leading to better results and generate more enthusiasmsinformation systems and numerous interoperability among the municipalities.problems. The concentration on a few projects can 7
  8. 8. Proceedings of the 39th Hawaii International Conference on System Sciences - 20064) Develop standard infrastructure components Information systems are built on top of a complex 8) Stimulate the formation of coalitionslayer of computer and software infrastructure to The agencies tend to do things on their own andcommunicate with other systems and making easy resist the initiatives of the national levels. Coalitionsaccess to the readily available infrastructure service. having participants from multiple agencies at aOver time a number of basic infrastructure components national and local level might breed new ideas.evolved that were likely to be used by every agencyand had limited variety. The reuse of readily available 6. Discussionand proven infrastructure components can help todevelop new systems quicker. 6.1. Usability of CAS for enterprise architects5) Develop modular architectures By conceptualizing enterprise architecture as a Whenever heterogeneous types of systems were CAS, public managers can interpret enterprisedeveloped, there were inevitable giving rise to architecture in a more complete manner and developinteroperability problems. Instead of being involved in more effective interventions. From our case studies, itthe design of complex systems architectures, the idea becomes clear that CAS offers an alternative,of this principle is to define the basic component interpretative framework to conceptualize thefunctionality and interfaces to invoke the components. behaviors among actors within the environment ofThis should constrain the variety of systems and ensure enterprise architecture. Awareness of the architecturalthat the systems can interoperate with each other. principles can help public administrators and managers to target resources for more effective interventions.5.3. Increase cross-boundary and level To most of the enterprise architects, the principlesinteractions derived might not be come as a surprise. However, what should be noted is that despite the fact that these CAS is based on the concept that the ways in which principles might be known and intuitively appealing,the agents in a system connect and relate to one most of the initiatives of enterprise architecture haveanother is critical to the survival of the system and is failed to implement these design principles. Ourmore important than the agents themselves. As such, objective here is to pinpoint the CAS not only can beenterprise architecture should provide interaction used to facilitate conceptualization and betterpoints that foster the building of relationships among understanding of the underlying dynamics but also canagencies. offer the design architects the theoretical underpinnings and explanation to support the selection6) Stimulate sharing of design choices and solutions. This can help avoid The municipalities are free to design their own interventions, which may lead to undesired behavior.systems and choose the means to accomplish this.Municipalities started looking for innovative ways to 6.2. Balancing central and decentral layersaccomplish the targets set at the national level, andafter budgets were cut, targets could not be met on The networks of public agencies exhibit local astheir own. Sharing of ICT-departments, functionalities, well as global complexity. The decision makingservices at a local level helps to reduce costs, also occurring at the local level is very different from thatincrease the available budgets in this way providing occurring at the global level and hence theaccess to expertise and systems formerly out-of-reach. complexities as well as its measures are different. At a local level other objectives are important than at the7) Develop competencies global level. The global level aims at balancing the If people at the local level have not the overall interests of the many different local agencies.competencies to fulfill a particular task, they are The decisions about which functions should beunable to adopt the results of national projects or meet performed at the centralized and decentralized level istargets. Knowledge and capabilities are necessary to a critical decision at a strategic level. Such a decisionuse and integrate the infrastructure components and has a long-term impact can have significantother results of the ICTU program. There should be complexity and risks. Peak and Azadmanesh’s [24]mechanisms in place to develop those capabilities. research suggests that strategies oscillate unevenlyPrior to starting something the necessary competencies between domination of centralization andshould be identified by organizing seminars, starting decentralization, and that commercial computing hascourses and so on. already experienced two centralization/decentralization 8
  9. 9. Proceedings of the 39th Hawaii International Conference on System Sciences - 2006cycles. They observed that centralization emphasis initiatives. We found that the CAS lens can help toshifts based on changes in technology. Therefore there facilitate conceptualization and better understanding ofis a need for constant re-assessment of the underlying dynamics. Moreover, CAS theory offerscentralization/decentralization options. enterprise architects the theoretical underpinnings and Designing central infrastructure developed only explanation to support the selection of design choicesslowly as each organization is autonomous and and solutions.controls its own technical architecture and decisions. Although we derived the architectural designWhat seems to work is to wait for the emergence of principles based on interviews using CAS as abasic infrastructure components at the local level, metaphor. We did not validate the application of thesegeneralize and scale them up. Despite this success, the principles in new projects nor did we develop a modeltranslation of policy into implementations of the basic of the complex adaptive system. Future work shouldinfrastructure components discussed in our case studies focus on evaluating the implications of principles ontook in average five years. complex-adaptive systems. A multi-agent simulation The development of modular functionality and model might be used, as agents can be used to modeldefining of interfaces to guide systems development the behavior of each individual entity and theseems to be suitable for managing the variety of aggregate behavior of individuals can show thesystems and overcoming the interoperability problem. dynamic and emergent behavior over time (e.g. [21],Determining the granularity of modules, ability to [26]). Once an agent-based model is created, variousconfigure and interfaces is no sinecure and more easily parameters can be manipulated and rules could bestated than performed. In the past, they tried to define modified in order to study the emergent outcomes. Inthe functionality and relationships among them at the this way the implications of design principles can benational level, which resulted in confusions and lack of evaluated and used to modify the design principles.adoption. Here, it also seems to be of importance thatsuch system architectures emerge from the local level, 8. Referencesinstead of trying to define them at a national level. [1] Armour, F.J., Kaisler, S.H. “Enterprise Architectures:7. Conclusions Agile transition and Implementation”, IEEE IT Professional, Vol. 2, No. 6, 2001, pp. 30-37. In this paper we developed a broader understanding [2] Auyang, S.Y., Foundations of Complex-Systemof use of enterprise architectures in the Dutch public Theories in Economics, Evolutionary Biology, andadministration by analyzing the key interaction points Statistical Physics. Cambridge University Press, Newfrom a complex adaptive systems (CAS) perspective. York.A complex system consists of a network of highly [3] Axelrod, R.M. The Complexity of Cooperation.interactive and interdependent entities. Rather than Princtone University Press, NJ, 1997.being planned or controlled the entities organize [4] Bakos, J.Y. “A Strategic Analysis of Electronicthemselves into patterns to be adaptive to its Marketplaces”, MIS Quarterly, vol. 15, no. 3, 1991, pp.environment. From a CAS perspective, a number of 294-310.relatively simple principles can guide development and [5] Boxtel, R. et al. Nationaal actieprogrammaa key challenge is to create an enterprise architecture elektronische snelwegen. Report Ministry of Interior,that merely sets the direction for the future, ensures 1998.uptake of emerging initiatives, and is ready to adapt [6] Boxtel, R. et al. De digital Delta. Report Ministry ofand evolve as the environment changes. This ideas Interior, 1999matches reality as tight control is not something you http://www.elo.nl/elo/Images/DDD_tcm70-16126.pdf,.can have over a complex system of agencies and [7] Bouwman, H. & Versteegt, G. “Business Architecture:projects spanning multiple levels and jurisdictions. A New Paradigm To Relate Business Strategy To ICT”, We investigated 11 e-government projects and European Conference on Information Systems, 2004.analyzed the interactions between agencies at a [8] Bouwman, H., Versteeg, G., Janssen, M., Wagenaar,national and local level. We derived architectural R.W. “From Business to services architectures:design principles using complex adaptive systems concepts and a research agenda”, Report Delfttheory that can be used by enterprise architects to University of Technology, 2005.direct further projects. They are based on the primarily [9] Choi, T.Y., Dooley, K.J. & Rungtusanatham, M.findings that solely breeding of diversity is not “Supply Networks and complex adaptive systems:sufficient, there is a need to have a focal point to control versus emergence”, Journal of Operationsconcentrate activities and efforts and an increase in the Management, Vol. 19, 2001, pp. 351-366.number of interactions result in new and better 9
  10. 10. Proceedings of the 39th Hawaii International Conference on System Sciences - 2006[10] Desai, A. ”Adaptive complex enterprise”, n/Pechtold+brengt+bureaucratie+voor+burgers+in+kaar Communications of the ACM, vol. 48, No. 4, 2005, pp. t/, 2005. 32-35. [26] Railsback, S.F. “Concepts from complex adaptive[11] Eisenhardt, K.M. “Building theories from case study systems as a framework for individual-based modeling: research”, Academy of Management Review, vol. 14, Ecological Modelling, vol., 139, 2001, pp. 47-62. No. 4, 1989, pp.532-550. [27] Ramnath, R., & Landsbergen, D. ”IT-enabled sense-and[12] Fayad, M.E., Hamu, D.S. and Brugali, D. “Enterprise –respond strategic in complex public organizations”, Frameworks Characteristics, Criteria, and Challenges”. Communications of the ACM, Vol. 48, no. 5, 2005, pp. Communications of the ACM, Vol. 43, No. 10, 2000pp. 58-64. 39-46. [28] Reschenthalter, G.B. & Thompson, F. “The information[13] Fryer, P. What are Complex Adaptive systems? revolution and the new public management”, Journal of http://www.trojanmice.com/articles/complexadaptivesys Public Administration Research and Theory, vol. 6, no. tems.htm [last accessed 11-6-05]. 1, 1996, pp. 125-143.[14] Goodhue, D.L., Kirsch, L.J., Quillard, J.A. & Wybo, [29] Richardson, G.L., Jackson, B.M. & G. Dickson, “A M.D., “Strategic Data Planning: Lessons form the Principle-Based Enterprise Architecture: Lessons From field”, MIS Quarterly, vol. 16, no.2, 1992, pp. 11-34. Texaco and Star Enterprise”, MIS Quarterly, Vol. 14,[15] Graaf, T. De Andere Overheid. Report Ministry of the No. 4, 1990, pp. 385. Interior and Kingdom Relations, [30] Rohloff, M. “Enterprise architecture: Framework and www.andereoverheid.nl, 2004. methodology for the design of architecture in the large”,[16] Hackney, R.A. & McBride, N.K. ‘The efficacy of European Conference on Information Systems, 2005. information systems in the public sector: issues of [31] Simon, H. A. “The architecture of complexity”, Proc. context and culture”, International Journal of Public American Philosophy Society, vol. 106, no. 6, 1962, Sector management, Vol. 8, No. 6, 1995, pp. 17-29. 467-482[17] Hazlett, S.A. & Hill, F. “E-government: the realities of [32] Simon, H.A. The sciences of the Artificial, MIT Press, using IT to transform the public sector”, Managing Cambridge, MA, 1969. Service Quality, Vol. 13, No. 6, 2003, pp. 445-452. [33] Smolander, K., Rossi, M. & Purao, S. “Going beyond[18] Holland, J.H. Hidden order. How adaptation creates the blueprint; Unraveling the complex reality of complexity. Addison Wesley, Reading, MA, 1996. software architectures”, European Conference on[19] Jaeger, P.T. “The Endless Wire: E-government as Information Systems, 2005. global phenomenon”, Government Information [34] Tan, C.W. & Pan, S.L. “Managing e-transformation in Quarterly, Vol. 20, 2003, pp. 323-331. the public sector: an e-government study of the Inland[20] Janssen, M. & Cresswell, A. “An Enterprise Application Revenue Authority of Singapore”, European Journal of Integration Methodology for E-Government”, Journal Information Systems, Vol. 12, 2003, pp. 269-281. of Enterprise Information Management, Vol. 18, no. 5, [35] Yin, R.K., Case Study Research: Design and methods. 2005 (forthcoming). Sage publications, Newbury Park, California, 1989.[21] Janssen, M. & Verbraeck, A. “An Agent-based Simulation Testbed for Evaluating Internet-based Matching Mechanisms”, Simulation Modelling Practice and Theory, Vol. 13, No.5, 2005, pp. 371-388.[22] Lewin, R. & Regine, B. The Soul at Work: Complexity Theory and Business, As If People Matter. Simon & Schuster, 1999.[23] Ongaro, E. “Process management in the public sector: The experience of one-stop shops in Italy”, International Journal of Public Sector Management, Vol.17, No. 1. 2004, pp. 81-107.[24] Peak, D.A. and Azadmanesh, M.H., “Centralization /Decentralization cycles in computing: Market evidence”, Information & Management, 31, 1997, 303- 317.[25] Pechtold, A. Resultaten van de onderzoeken naar administratieve lasten. Letter to the parliament of the Minister of Innovation, http://www.lastvandeoverheid.nl/LVDO/Web/Resultate 10