Swarnendu Biswas
Student: MS (by research)Roll: 09CS7001
Department of Computer Science and Engineering
IIT Kharagpur - 721302
| |
Swarnendu BiswasStudent: MS (by research)Roll: 09CS7001 Department of Computer Science and Engineering IIT Kharagpur - 721302 |
|
Room 206, GM CRL,
Department of CSE,
IIT Kharagpur,
Kharagpur - 721302.
Phone - (91) - 32222 - 81442
Email: swarnendu [AT] cse.iitkgp.ernet.in
My areas of interest are Software Engineering, Embedded Systems, Real-Time Systems and Compilers.
My current MS. work is on Regression Testing of Embedded Programs. My supervisor is Prof. Rajib Mall. Here is a brief summary of my work.
Existing regression test selection techniques often fail to select all potentially fault-revealing test cases when applied to embedded programs. This is because these techniques ignore many important embedded program semantics during selection of test cases, e.g., tasks, timers, inter-task communication, etc. In this context, we have proposed a model for embedded programs by extending the System Dependence Graph (SDG). In addition to control and data dependence information, our SDGC model can represent important features of an embedded program such as tasks, task priorities, precedence relationships, timer constructs, and inter-task communication using message queues and semaphores. We have proposed a regression test selection technique based on an analysis of these models. In addition to data and control dependences, our technique also takes into account the timing dependences introduced among tasks in an embedded program due to precedence order relations, task priorities, inter-task communications and interrupt handlers.
In an industry scenario, it is often not possible to execute all the relevant test cases due to constraints on resources for regression testing. Regression testing is usually contrained by restriction on availability of resources such as time, budget, personnel, etc. In this context, we propose a multi-objective regression test suite optimization technique for embedded programs. Our optimization technique identifies tasks that need to be tested because they may potentially be affected due to the timing dependencies that are introduced among tasks in an embedded program due to precedence order and inter-process communication using message passing and shared memory. Test cases which execute critical functionalities and tasks whose timing behavior may potentially be affected are not omitted during optimization. Additionally, our technique also aims to minimize the cost of regression testing, maximize the reliability of the frequently-executed functionalities and remove redundant test cases.
I am working as a Research Consultant at IIT Kharagpur on
projects sponsored by General Motors India since September, 2008. My work is on automated and effecientregression testing approaches for automotive software.
I have worked for more than three years with Wipro Technologies as a project engineer across locations such as Pune and Kolkata. I was involved in sofware development in the Unified Communications and Automotive Software domain.
These projects have been carried out duing my tenure at IIT Kharagpur (September 2008 - present) in the capacity of a Research Consultant.