Mounia Fredj

Component-Based Development: Extension with Business Component reuse

Component-Based Development Methods (CBDM) become largely adopted and used. They replace the traditional methods by the introduction, in their development lifecycle, of two processes: design for reuse and design by reuse. However, these two processes are not clearly established and are combined in the majority of the CBDM. This paper aims to propose an approach to extend a CBDM by the reuse processes, using the Symphony method as a reference. A special interest is given to the reuse of Business Component for which we propose a conceptual model based on the variability concept that we consider a key property in a reuse process.

A pattern based approach for requirements engineering

This paper presents a pattern-oriented approach to requirements engineering. It proposes an OrganizationDiagnosis pattern language which attempts to represent the fundamental elements intervening recurrently during requirements engineering elicitation. The idea is to propose components that enable to reuse the solutions for recurrent problems to which analysts are confronted in software engineering. Therefore, we propose to capitalize the now-how in requirements engineering by adopting a pattern based approach. The analyst can select components that he judges relevant for its application and adapt them to its context. Each pattern describes the problem, its context and the solution in a semi-formal language. Our approach identifies a set of patterns that can serve as nucleus to elaborate the system analysis. An illustration of the OrganizationDiagnosis pattern language is given through an application of order management.

To trace and guide evolution in configurable process models

Today, research on business process management (BPM) outlines the importance of business process reuse through reference process models which are interested in capturing common practices of organizations. Consequently, reusing these reference process models stresses the necessity of managing a set of process variants. In this context, configurable process model solutions have been developed to facilitate customization of business process in a specific scenario. However, configurable process model should evolve in order to meet new business requirements. To address this issue, we propose a process patterns based approach for leading configurable process model evolution. For supporting traceability of the transition from a given version of a configurable process model to another, an evolution metamodel has been defined. In this paper, we present i) only one pattern of our proposed patterns system, the Activity substitution process pattern, and ii) a proposition of the evolution metamodel.

A process pattern for managing evolution of configurable reference process models

During the last years, reference process modeling becomes an efficient approach for reuse in business process modeling (BPM). Many BPM tools provide mechanisms to adapt reference process models in order to meet new needs. However, this adaptation results in a set of variants which have to be managed. One of the most proposed solutions is the configurable process model. Therefore, evolving this configurable process model in terms of activities, a data flows and resources changes becomes a crucial issue. In this paper, we propose a process pattern which guides designers for managing evolution of configurable reference process models.

Support for resources changes in configurable process models

Today, researches on business process models reuse are interested in configurable process modeling which is an important approach for reusing and customizing business process models with a set of process variants. However, enterprises are involved in a competitive environment which imposes them to often evolve their configurable process models in terms of activity, data and resource. So, creating configurable process models becomes insufficient since the configurable process models should evolve to meet new needs. To address this issue, we propose a reusable process patterns system for guiding designers in managing changes and evolution of configurable process models. In this paper, we focus on how to manage resources changes. We present one pattern of our proposed system, the variation point resource insertion process pattern.

A Hybrid Strategy for Integrating Sensor Information

The combination of sensor networks with databases has led to a large amount of real-time data to be managed, and this trend will still increase in the next coming years. With this data explosion, current integration systems have to adapt. One of the main challenges is the integration of information coming from autonomously deployed sensor networks, with different geographical scales, but also with the combination of such information with other sources, such as legacy systems. Two main approaches for integrating sensor information are generally used: virtual and warehousing approaches. In the virtual approach, sensor devices are considered as data sources and data are managed locally. In contrast, in the warehousing approach, sensor data are stored in a central database and queries are performed on it. However, these solutions turn out to be difficult to exploit in the current technology landscape. This paper focuses on the issue of integrating multiple heterogeneous sensor information and put forward a framework for decision making process.

Metamodels for models complete integration

Problems occur in the models integration and particularly, in the comparison phase of the integration. Previous approaches have not correctly handled the semantic integration. We propose a complete approach for integrating of UML models. First, we provide a comparison approach incorporating syntactic, structural and semantic aspects in order to detect any correspondence between the models to compare, and secondly, a semantic approach to integrate the compared models. For this purpose, we use domain ontology. In this paper, our contribution is the proposition of the different metamodels needed in the comparison and the integration phases.

Variability Management in Business-IT Alignment: MDA based Approach

The expansion of PAIS (Process Aware Information Systems) has created the need for reuse in business processes. In fact, companies are left with directories containing several variants of the same business processes, which differ according to their application context. Consequently, the development of PAIS has become increasingly expensive. Therefore, research in business process management domain introduced the concept of configurable process, with the aim of managing the variability of business process. However, with the emergence of the services-based development paradigm, the alignment of services with business processes is highly required in PAIS. Thus, in this paper an MDA based method which allows for generating configurable services from configurable process is proposed.

Integrating linked sensor data for on-line analytical processing on-the-fly

Sensor networks are gaining more and more attention in the current technology landscape. It is undeniable that their use allows a better monitoring of events that occur in the real world. Many sensors have been deployed for monitoring applications such as environmental monitoring, and traffic monitoring. A number of governments, corporates, and academic organizations or agencies hold independently sensor systems that generate a large amount of dynamic information from data sources with various formats of schemas and data. They are making this sensor data openly accessible by publishing it as Linked Sensor Data (LSD) on the Linked Open Data (LOD) cloud. LSD is the concept that defines the publication of public or private organization sensor data without restrictions. This is achieved by transforming raw sensor observations to RDF format and by linking it with other datasets on the LOD cloud. The seamless integration of LSD sources from multiple providers is a great challenge. In this paper, we investigate the possibility of integrating diverse LSD sources using the hybrid ontology approach for on-line analytical processing (OLAP) on-the-fly. With such an ontology-based integration framework, organizations or individuals will have greater opportunity to make their respective analysis based on a large amount of sensor data openly accessible on the Web.