Shopping Cart
Your Cart is Empty
There was an error with PayPalClick here to try again
CelebrateThank you for your business!You should be receiving an order confirmation from Paypal shortly.Exit Shopping Cart

Testing and Coverage Visualization of Data-intensive Systems

Our current work focuses on testing very large data-intensive systems used in e-governance. This high-impact project aims to help the Norwegian Toll Customs test, verify, and validate their software system EMIL. EMIL is a central script that processes EDIFACT messages representing customs transactions to update the national database of goods entering Norway. About 30,000 messages are processed every day. Our role in this project is to transfer technology including pairwise coverage and coverage visualization to help testers at Toll Customs realize their regression testing.


1. Sen S., Gottlieb A.,Testing a Data-intensive System with Generated Data Interactions: The Norwegian Customs and Excise Case Study, CAiSE 2013 [PDF]

Self-adaptive and Context-aware Software Systems

We are exploring novel software architectures based on “dynamic component loading” to develop self-adaptive software. In particular, we have built self-adaptive vision systems that can change behavior at runtime given a change in context.

An Example Our Work

In the right handside Fig. (a) , we present the architecture of Tekio. The architecture in spirit is based on the MAPE-K loop from IBM. A Context Manager receives events from software components (encapsulating application-specific behavior, hardware sensors such as cameras, etc.) and produces a context configuration. The Adaptation Planner computes a system configuration that is an appropriate answer to the give context configuration given a set of rules. The Component Executer reconfigures the system with the new configuration. In Fig. (b), we show that a system running in the intrusion detection mode (for usually a long period of time with lower resource requirements) switches to face detection Fig. (c) (requiring more resources) only when movement is detected. This scenario presents a sustainable solution to the video surveillance problem that records hours of static videos with seldom occurrences of movement.

Selected Publications

1. Hurtado Santiago, Sen S., Cassallas Rubby,Reusing Legacy Software in a Self-adaptive Middleware Framework, Adaptive and Reflective Middleware Workshop, Middleware 2011 [PDF][BibTex]

2. Rocha Leonardo, Sen S., Moisan Sabine, Rigault Jean-Paul Girgit: A Dynamically Adaptive Vision System for Scene Understanding, International Conference on Vision Systems, Sophia-Antipolis, France 2011 (20% Acceptance Rate) [PDF][BibTex]

3. Rocha Leonardo, Sen S., Moisan Sabine, Rigault Jean-Paul, Girgit: A Framework for Component-based Dynamic Adaptation in Python, EuroSciPy 2011, Paris, France (Demo)

Design Space Exploration in Model-driven Engineering

One of my primary research interests is to elucidate ways in which human creativity, knowledge from various sources and brute computational power can go hand in hand to synthesize useful and complex software systems. This is the problem of design space exploration given heterogeneous sources of knowledge.

Humans must do what they are good at (such as provide intuitive opinion, integration of knowledge, creative power and strategies to address scalability) while computers must do what they do well (such as large-scale automatic constraint satisfaction). How can I systematically combine various sources of knowledge to automatically explore the design space of high-level descriptions of software systems called models? This is the question that intrigues me.

My doctoral work entitled “Automatic Effective Model Discovery” was a practical step in this direction. My work focusses on scaling formal methods to perform guided search in the design space of high-level descriptions of software called models. The principal contributions of this work include partial model completion in domain-specific model editors and test model generation.

An Example of Our Work

In Fig. below, we present a set of automatically discovered test models using our tool Pramana that was used to detect bugs in a model transformation from Class Diagrams to RDBMS.

Selected Publications

1. Mottu J-M, Sen S., Tisi M., Cabot J., Static Analysis of Model Transformations for Effective Test Generation, ISSRE 2012, [PDF][BibTex]

2. Sen S., B. Baudry, Vangheluwe H., Towards Domain-specific Model Editors with Automatic Model Completion, Simulation : Transactions of the Modeling and Simulation Society, 2009 [PDF] [BibTex] [Final SAGE Version]

3. Sen, S., N. Moha, B. Baudry, and J.-M.Jezequel, Meta-model Pruning. In Proceedings of MoDeLs. 2009. Denver, Colorado. (Acceptance rate: 16%) 2009 [PDF] [BibTex]

Product Discovery in Software Product Lines

A software product line is the set of variants of a software system. The AspectOptima feature model shown in the Fig. (left) is one way to represent the variations. Some of these variants can be optimal for a given task. Discovery of products for testing and validating a large software system is what intrigues us. This project forms the basis for the our work on web service orchestrations.

Selected Publications

1. Perrouin Gilles., Oster Sebastian. Sen S., Baudry Benoit, and Le Traon Yves, Pairwise Testing for Software Product Lines: A Comparison of Two Approaches, Software Quality Journal 2011. [PDF] [BibTex]

2. G. Perrouin, Sen S., Klein J., Baudry B, Le Traon Yves, Automated and Scalable T-wise Test Case Generation Strategies for Software Product Lines, ICST 2010, Paris, France (Acceptance rate 26.5%) [PDF] [Bibtex]

Exploration and Validation of Web Service Orchestrations

Web services can be orchestrated in many ways to realize different composite services. We explore this constrained space of variations typically to evaluate design alternatives and validate a composite web-service. The evaluation is based on executing web services with offline models of QoS such as probability distributions.

An Example of Our Work

In Fig. (a) on the left we present an example offline model of a probability distribution for an atomic web service. While in Fig. (b) we display the variations we observed in a composite web-service for e-Health based on the a composition of the atomic services such as ambulance, hospital records, etc.

Selected Publications

1. Kattepur Ajay, Sen S., Baudry B., Beneveniste A., Jard C. Pairwise Testing of Dynamic Composite Services. (SEAMS 2011 co-located with ICSE 2011) [PDF] [BibTex]

2. Kattepur Ajay, Sen S., Baudry B., Beneveniste A., Jard C. Variability Modeling and QoS Analysis of Web Services Orchestrations. (ICWS 2010, @17.6%) [PDF] [BibTex]

Research Projects

I am or I have been part of a certain number of projects that graciously fund my research activities.

Toll Customs, Norway

I am currently (since September 2012-) involved in a project with the Norwegian Toll Customs department. Our goal is to develop tools to help testers visualize coverage of their test cases in a very large databse system used for E-governance.

S-Cube Network of Excellence

I was involved in the activities of the European S-Cube network of excellence project for about two years 2008-2010. My task involved attending some project meetings, publishing articles in the area of service-oriented computing, and contributing to deliverables.


In the year 2007-2008, I was involved with the Agence Nationale de Recherche (ANR) funded project called DOMINO. My responsibilities included developing techniques for model transformation testing for modelling domains available at CNES and Airbus. A number of meetings were held across France to present our work. An example of my research involved generating UML activity diagrams to test a transformation from UML Activity Diagrams to embedded code on a star tracker docking system for CNES.