R&D Engineer in Cluster Operating Systems

Place: INRIA Rennes - Bretagne Atlantique Research Centre (France)
http://www.irisa.fr/ 
http://www.inria.fr

Team: PARIS project-team (http://www.irisa.fr/paris)

Duration: Initial contract for 12 months + contract renewal up to May 2010

Starting date: as soon as possible

Context
    This position is proposed in the framework of the XtreemOS European FP6 Integrated Project (http://www.xtreemos.eu). XtreemOS project aims at building and promoting a Linux-based Operating System to Support Virtual Organizations for Next Generation Grids. INRIA is the scientific coordinator of XtreemOS project. The consortium is composed of 19 industrial and academic partners from 8 different countries (7 European countries and China).
     
The R&D engineer will participate in the research and development activities related to the design and implementation of a Grid operating system that will be conducted by INRIA Rennes within the XtreemOS project.

Work description
The research and development engineer will participate in the design and implementation of XtreemOS, a Grid system based on Linux operating system.

She/he will be in charge of the activities on LinuxSSI cluster operating system INRIA is responsible for in the framework of XtreemOS. LinuxSSI is the foundation layer of the XtreemOS Grid operating system for cluster nodes. The Grid level XtreemOS services are run on top of LinuxSSI. LinuxSSI is a single system image operating system leveraging the Linux based Kerrighed cluster operating system, originally developed by INRIA in collaboration with EDF R&D and now evolving as an open source software (http://wxww.kerrighed.org). LinuxSSI gives the illusion that a cluster is an SMP machine.

The engineer will contribute to the design and implementation of LinuxSSI, to its integration in the XtreemOS cluster flavour, to its packaging in Linux distributions and will participate in supporting the user community. In particular, she/he will participate in the following tasks:

-    Design and implementation of advanced features for the KDFS distributed file system which is part of LinuxSSI, in particular reconfiguration mechanisms to support node addition & eviction,
-    Design and implementation of kernel level  mechanisms integrated in LinuxSSI to checkpoint applications on cluster Grid nodes (improving Kerrighed process checkpointer to meet XtreemOS needs and extending it to processes communicating through messages),
- Testing, experimentation and validation of the implemented mechanisms including performance measurement.
   

In addition to the software design and development, the engineer will write documentations (installation, user, reference manuals) and contribute to the deliverables (reports and prototypes). She/he will participate to technical meetings and to the promotion of the XtreemOS software (preparing and presenting demonstrations, posters, tutorials).
  
Experimentations will be carried out on clusters from the Aladdin-G5K experimental Grid platform (http://www.grid5000.fr) consisting of PC clusters located in 9 different sites (INRIA – Rennes being one of the Aladdin-G5K sites with more than 500 processors).

The engineer will interact with engineers, researchers and PhD students working on the design and implementation of operating systems for clusters and Grids within the PARIS project-team at INRIA Rennes. She/he will also interact with the other industrial and academic partners of the XtreemOS consortium for the XtreemOS development and experimentation (in particular its cluster version based on LinuxSSI). She/he will also interact with developers and users from the Kerrighed open source community.

Profile
    PhD degree in Computer Science or engineering or master degree in Computer Science + experience

Required expertise
-    Good knowledge in operating systems and distributed systems design and implementation
-    Knowledge of Linux kernel (experience in Linux kernel module development would be very much appreciated)
-    Good knowledge of C and Java programming language
-    Good knowledge in distributed file systems
-    Experience with Linux and open source software projects would be very much appreciated
-    Experience in software development in a large project would be appreciated
-    Very good knowledge of English (read, spoken, written) demanded
-    Ability and liking for team work and relationships
-    Rigor, autonomy, organizational skills
-    Good writing ability

Contact
Christine Morin
XtreemOS project leader
PARIS project-team
IRISA/INRIA
Campus universitaire de Beaulieu
35042 Rennes cedex (France)
Email: Christine.Morin(@)inria.fr
Phone: +33 2 99 84 72 90

Application
Applicants should send their application file including a motivation letter, a full resume and up to five reference letters to
Sandrine L’Hermitte
INRIA Rennes – Bretagne Atlantique
Campus de Beaulieu
35042 Rennes cedex (France)
Email: Sandrine.L_Hermitte(@)irisa.fr