Running software in uncertain environments: how diversification and approximation can enhance robustness
Dr. Benoit Baudry from INRIA, France
Tid: Fr 2016-10-28 kl 13.30
Plats: D1, Lindstedtsvägen 17, KTH
Software applications run in environments that are getting more complex every day: hardware is increasingly approximate, users want more customization, applications rely on third party modules that are most often running remotely and hackers intensify their attacks.
Consequently, software behavior needs to be more adaptable and flexible to cope with perturbations in its environment.
In this talk I will survey works that have investigated how software can be made more approximate and diverse in order to address this challenge.
These works range from approximate source code transformations to runtime fault injection. They are grounded in the areas of compilation, programming languages and program transformations and are most often inspired by natural sciences such as immunology and ecology.
Benoit Baudry is a research scientist at INRIA, France. His research is in the area of software testing and analysis. He currently works on automatic software diversification and test generation for the construction of robust systems, in tight collaboration with companies such as Orange, ATOS and SMEs. He leads the DiverSE research group, a software engineering group that investigates automatic composition and synthesis of software diversity to manage unpredictability. He coordinated the FP7-FET DIVERSIFY project (2013-2016) and coordinates the STAMP H2020 project (2016-2019). He received his PhD from the University of Rennes in 2003.