"Efficient interactive workflows for scientific applications"
Starting in: september 2008 (at the latest january 2009)
Duration: 3 years (6 up to 12 months extension possible)
Salary: 1900 Euro/month pre-tax (1500 Euro/month net)
INRIA is the leading French research institute in computer science. Its Saclay center, half an hour away from Paris, has been created in january 2008 by a split of the former Futurs center in three centers (Saclay, Bordeaux, and Lille). The Saclay center benefits from an unique environment bringing together the largest science university in France, including the prestigious CS research laboratory LRI, higher-education institutes like Ecole Polytechnique and Supelec, and industrial research labs such as the INRIA-Microsoft center, Thales, CEA etc.
Topic description
The proposed topic bridges three related yet distinct areas: (1) workflow modeling; (2) database execution and optimization; and (3) information visualization. The thesis will be carried in a collaborative environment in the south of Paris, bringing together internationally recognized researchers in these areas.
Context
Today's scientific data management applications involve huge and increasing data volumes. The availability of such large data volumes is a gold mine for scientists who may carry research at an unprecedented scale with this data. Today's scientists, however, more often than not rely on proprietary, ad-hoc information systems, e.g., a directory structure organized by hand by the scientist, a few specialized data processing applications and perhaps a few scripts etc. Significant work has been invested in the development of workflow platforms. However, their models do not adequately take into account intensive user intervention at any level and scale during the execution of a process. Efficient deployment on large data volumes (while preserving flexibility) is another concern.
Envisioned approach
The purpose of the thesis is to investigate models, algorithms, and propose an architecture of an interaction- and data-intensive workflow engine adapted to the task:
(1) A flexible yet efficient generic data model will be chosen, inspired from existing formats currently in use in scientific applications, like HDF, NetCDF etc.
(2) A process model in the style of data flow must be proposed, which would allow the user to inspect, interrupt, resume or guide the process with little effort at any point.
(3) A workflow platform will be prototyped based on existing open-source components; database-style optimization techniques should be developed in this context.
Required knowledge and background
The student should be familiar with database technologies and knowledgeable with visualization or workflow systems.
To apply
- INRIA official recruitment web site: http://tinyurl.com/2cjqbh
Follow the link "Apply online" at the bottom of the page
- Contact:
Jean-Daniel Fekete, Aviz Team, INRIA Jean-Daniel.Fekete@inria.fr
Ioana Manolescu, Gemo Team, INRIA Ioana.Manolescu@inria.fr
Veronique Benzaken, LRI, U. Paris Sud Veronique.Benzaken@inria.fr
- Hard deadline for applications: may 15, 2008
- Notification of acceptance: june 10, 2008
French students having completed their MS in the Paris area are not eligible! Priority will be given to students with study/work experience outside France.
More information