Available online at www.sciencedirect.com
ScienceDirect Procedia Computer Science 64 (2015) 832 – 837
Conference on ENTERprise Information Systems / International Conference on Project MANagement / Conference on Health and Social Care Information Systems and Technologies, CENTERIS / ProjMAN / HCist 2015 October 7-9, 2015
Approach based on web services for business process adaptation Afef Awadida, Sonia Ayachi Gnannouchia* a
RIADI Laboratory, ENSI Manouba University,2010 Manouba, Tunisia
Abstract Business process lifecycle is chiefly divided into two phases namely modeling and execution. Therefore, flexibility of business process can be taken into account either at build-time or at runtime. The first kind is well supported mainly by declarative modeling paradigms. Nevertheless, many challenges are posed by the second kind of flexibility (such as at which level of abstraction can flexibility be expressed and by what criteria should it be performed) as well as by the business process adaptation when we consider the implementation phase of business process. In this context, our aim in this paper is to propose an approach based on web services in order to support business process adaptation during its implementation. The proposal is to implement as web services the manners in which activities are executed and that can be subject to change. Each manner refers to the actions implementing the current activity. It can then be adapted to execute the same business process in another organization. © 2015 2015The TheAuthors. Authors. Published Elsevier © Published by by Elsevier B.V.B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). Peer-review under responsibility of SciKA - Association for Promotion and Dissemination of Scientific Knowledge. Peer-review under responsibility of SciKA - Association for Promotion and Dissemination of Scientific Knowledge Keywords: Business process instances; flexibility by adaptation;web service;imperative modeling paradigms
1. Introduction Nowadays, companies face high rates of change from a large number of sources. Inside pressures come from human actors who push for change. Outside pressures come from changes in the competitive, technological and economic environments. These changes are considered as essentials for companies in order to remain competitive, better meet customer needs, enhance internal processes and modify the range of their products or services1. In fact, changes of different natures can strongly affect business processes. This plethora of changes may occur due to regulations, appearance of new competitors and technological evolution. Accordingly, organizations constantly have to adapt their processes.
*
E-mail address:
[email protected]
1877-0509 © 2015 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). Peer-review under responsibility of SciKA - Association for Promotion and Dissemination of Scientific Knowledge doi:10.1016/j.procs.2015.08.635
Afef Awadid and Sonia Ayachi Gnannouchi / Procedia Computer Science 64 (2015) 832 – 837
Depending on change’s abstraction level, each form of change generally concerns either process specification or process instance. However, when we consider the implementation level, we find that the manners in which process activities are executed (in other words the actions implementing process activities) may be also subject to change. Thereby, business processes in various fields such as those of higher education and insurance, often undergo changes related to decisions and regulations. In the sense that such organizations share practically the same process schema, while the actions implementing their process activities differ from one organization to another. Thus, considering the rigidity of imperative modeling paradigms such as Business Process Model and Notation (BPMN), we highlight the need for approaches supporting the business process adaptation during the implementation phase. In fact, this phase needs significant coding effort aiming at implementing actions executing process activities. In order to promote reuse of these actions by other organizations having the same process schema, we propose to implement them as web services. In this paper, we propose an approach consisting of adapting actions implementing process activities by relying on web services. The remainder of the paper is organized as follows. Section 2 is devoted to present and discuss related work. Section 3 proposes an approach based on web services for supporting business process adaptation. Section 4 illustrates the proposed approach with an example and section 5 is dedicated to concluding our paper. 2. Related work The study of the literature shows that a significant number of researchers have focused on business process flexibility issue2-9. The main classification of these works is the one proposed by Andonoff et al.10. It puts forward six categories of approaches as follows: approach by evolution, approach by case, approach by conception, approach by constraints, approach by intention and context and approach by protocols integration. The approach by evolution includes all works that deal with process schemas evolution3, with process schemas versioning8 and with process instances migration11. The approach by case4, 6 represents a paradigm originally proposed by Van Der Aalst et al.4. It revolves around the idea of enabling a specific execution for each process instance by allowing actors to manage and control this execution using the data used and produced by the instance. This approach is distinguished from other approaches by focusing on what is feasible rather than what is prescribed. As for the approach by conception5, each work included in this category adopts Siebert’s viewpoint, who considers that one of the ways in which adaptability can be supported is to avoid it by combining flexible execution mechanisms with powerful modeling techniques in order to reduce the need for adaptability12. In this sense, approach’s feasibility depends strongly on the effectiveness of modeling techniques. Approach by constraints6 differs principally from procedural approach in how process schema is defined. In fact, the first presents a process schema as a sequence of activities and a set of constraints monitoring the execution of these activities not simply as a sequence of activities as is the case with procedural approach. The fifth kind of approaches corresponds to the intention and context-driven approach2, 7. In fact, based on specific modeling paradigms, it has the potential to take into account situational factors that can affect the process during its execution. This approach proposes to highlight adaptation needs using two complementary mechanisms that are variability and context awareness. Finally, regarding the approach by protocols integration9, the main idea of this adaptation form consists in integrating, in a smooth and coherent manner, process and interaction protocols10. The latter associates human actors to process in order to better enhance process guidance. From the point of view of changes that may arise during business process lifecycle, a taxonomy of flexibility was proposed by Regev et al.13. This taxonomy considers principally three dimensions characterizing a given change. In fact, it distinguishes first the abstraction level of change which concerns the level at which change is applied in a business process. This dimension determinates whether the change affects the process specification (process schema) or the process instance. Second, the change object concerning all aspects of process that may be subject to change. Third, the change properties, for instance the degree of change, which may be partial in order to modify a part of the process or total in the aim of creating new process. Taking into account the first dimension of the current taxonomy which is the abstraction level of change, we find that all mentioned approaches focus on business process adaptation, where the subject to change is either process instance (runtime) or process schema (design time). To promote this flexibility, many approaches such as4, 14 rely on
833
834
Afef Awadid and Sonia Ayachi Gnannouchi / Procedia Computer Science 64 (2015) 832 – 837
the power of declarative modelling paradigms. In the sense that, thanks to these paradigms it is much easier to reach flexibility. However, when we consider imperative modelling paradigms (marked by their rigidity) and actions implementing process activities as subject to change, we note that it seems important to focus on this challenge.
3. Approach based on web services for business process adaptation This section is devoted to describe our proposed approach as well as presenting its interests and challenges. 3.1. Approach description The fact that many geographically distant organizations share the same process schema and differ only in the actions implemented in order to automate their processes, leads us to introduce web service as a fundamental concept on which our approach is based upon. In this regard, web services can propose simple functions (request/response) that are accessible over the internet (which promotes their reuse). In fact, the novelty of our work consists in focusing on business process adaptation neither at design phase (when changes affect process schemas) nor at execution phase (when changes affect process instances) as is the case with all related work, but at implementation phase (when changes affect actions implementing process activities). Hence, the proposal is to implement as web services all actions implementing process activities, which are subject to change from one organization to another. Then, we have just to adapt actions automating process activities in a such organization in order to implement those automating the same process activities in another one. In this sense, adapting business process refers to adapting actions implementing its activities. In this paper, we consider business processes that are modelled with BPMN (we are interested mainly in private business processes; those internal to a specific organization). Therefore, the proposed approach is applied in case the used modeling paradigm is imperative. In the sense that the activities to be executed are known in advance. Further, we suppose that each time a change is required, related to a given process activity, a new web service has to be implemented (it has the activity name as request parameter and the actions implementing this activity as response parameter as shown in Fig. 1). This implementation has to be decided during the implementation phase (after the design phase).
Fig. 1. Implementing a web service for each activity whose actions are subject to change.
When implementing the same process activities in another organization, we need just to import for each activity the web service that returns the actions implementing it. These actions are then adapted in order to implement the process at hand. Our proposed approach is shown in Fig. 2.
Afef Awadid and Sonia Ayachi Gnannouchi / Procedia Computer Science 64 (2015) 832 – 837
Fig. 2. The proposed approach for business process adaptation.
3.2. Approach’s interests and challenges The presented approach: x Aims to help imperative modeling paradigm cope with its rigidity (inflexibility) by promoting business process adaptation. x Promotes the reuse of actions implementing process activities (which are subject to change) through recourse to web services. Thus, it avoids implementing actions automating activities from scratch. x Takes into account the evolving nature of the business process since it is iterative. However, dealing with business process adaptation involves many challenges such as the difficulty of considering all upcoming adaptations of a business process. For instance, adapting the structure (e.g., control flow) of a process instance during execution. Further, in the present work we have considered as BPMN processes only the private business processes. Therefore, we have to prove the applicability of our approach to interface (public/abstract) processes (which represent the interactions between a private business process and another participant) as well as to collaboration processes (which depict the interactions between two or more business entities).
835
836
Afef Awadid and Sonia Ayachi Gnannouchi / Procedia Computer Science 64 (2015) 832 – 837
4. Example In order to evaluate our proposed approach, we have modelled a business process in the field of higher education. In fact, the Tunisian Higher Education Institutions share the same process schema (process specification). However, actions implementing their process activities differ from one institution to another. This process consists of selecting candidates to join postgraduate study at master level. It is shown in Fig. 3.
Fig. 3. Business process model of selecting candidates.
In the next step, the actions to be implemented and that are subject to change from one institution to another are those automating the activities ‘calculate scores’ and ‘notify candidates’. Since for each institution, there is a particular formula used for calculating candidates’ scores. Further, there is a particular way to notify candidates (by e-mail, Short Message Service (SMS)). Therefore, we implement two web services related to the activities ‘Calculate scores’ and ‘Notify candidates’ as shown in Fig.4.
Fig. 4. Web service implemented for the activity ‘Notify Candidates’.
These web services are then published in order to promote their reuse for the same process schema but in another Higher Education Institution. In fact, for each other institution, we import the published web services. This enables accessing to actions automating the previous process schema in order to adapt them to the new needs. 5. Conclusion In this paper, we presented an approach for supporting business process adaptation, when the subject to change is the actions implementing its activities. Therefore, we dealt with the adaptation of business process at implementation level. The presented approach is based on web services in order to promote the reuse of these actions for implementing the same process activities in another organization. It avoids then implementing actions from scratch. Thus, the originality of our approach can be summarized into two points; focusing on the actions implementing process activities as the subject to change and implementing them as web services in order to promote their reuse.
Afef Awadid and Sonia Ayachi Gnannouchi / Procedia Computer Science 64 (2015) 832 – 837
837
Our approach was evaluated through a simple business process. This leaves many opportunities for future work to prove the effectiveness of our approach also in the case of complex business models as in the field of insurance, as well as to consider other imperative modeling paradigms such as Unified Modeling Language (UML). Moreover, the focus will be given to process execution aspects such as handling exceptions. References 1. Nurcan S. Engineering, Enterprise and Information Systems Architecture: Concepts, Foundations and Methods. Accreditation to Supervise Research (HDR). Paris 1 Pantheon-Sorbonne University; 2012. 2. Nurcan S, Edme M.H. Intention Driven Modelling For Flexible Workflow Applications. International Journal on Business Process Management. Special Issue on Software Process Improvement and Practice. vol. 10. n˚4; 2005. 3. Reichert M, Rinderle S, Kreher U, Dadam P .Adaptive Process Management with ADEPT2. International Conference on Data Engineering: Demo Session. Tokyo.Japan. pp. 1113-1114; 2005. 4. Van der Aalst W, Weske M Grunbaner D. Case Handling: a New paradigm for Business Process Support. International Journal on Data Knowledge Engineering. vol. 53. n˚2. pp. 129-162; 2005. 5. Adams M, Ter Hofstede A, Edmond D, Van der Aalst W .Worklets: A Service-Oriented Implementation of Dynamic Flexibility in Workflows. International Conference on Cooperative Information Systems. Montpellier. France. pp. 291-308; 2006. 6. Van der Aalst W, Pesic M, Schonenberg H. Declaration Workflow: Balancing between Flexibility and Support. International Journal on Computer Science- Research and Development. vol. 23. n˚2. pp. 99-113; 2009. 7. Saidani O, Nurcan S. Context-Awareness for adequate business process modelling. International Conference on Research Challenges in Information Science. Fes. Morocco. pp. 177-186; 2009. 8. Chabane MA, Andonoff E, Bouzguenda L Bouaziz R. Multidimensional modelling of process versions. Journal of Information Systems Engineering. vol. 15. pp. 89-114; 2010. 9. Faure C, Andonoff E, Hanachi C, Sibertin-Blanc C, Salatge N .Flexibility of Crisis Management Process by Integrating Protocols. Journal of Information Systems Engineering. vol. 15. n˚3. pp.37-60; 2010. 10. Andonoff E, Hanachi C and Nurcan S .Enterprise Process Adaptation. Adaptation in all its forms. P.H. Lopisteguy. Cepadues Editions. vol.2. pp. 78-122; 2012. 11. Casati F, Ceri S, Pernici B, Pozzi G. Workflow Evolution. International Conference on the Entity Relationship Approach. Cottbus. Germany. pp. 438-455; 2006. 12. Siebert R. An Open Architecture for Adaptive Workflow Management Systems. International workshop on Issues and Applications of Database Technology (IADT’98). Berlin.; 1998. 13. Regev G, Soffer P, Schmidt R .Taxonomy of Flexibility in Business Processes. Seventh Workshop on Business Process Modelling. Development and Support In conjunction with CAiSE’06; 2006. 14. Faustmann G. Enforcement vs. Freedom of Action- An Integrated Approach to Flexible Workflow Enactment. Workshop on Adaptive Workflow Systems. Conference on CSCW. Seattle. USA; 1998.