Simulation for Large-Scale Distributed Computing Research
This tutorial will provide attendees with clear perspectives on the challenges for experimental research in the area of parallel and large-scale distributed computing, and on current technology for conducting experiments with real-world testbeds, emulated testbeds, or simulated testbeds. The first part of the tutorial will present and contrast current experimental methodologies, giving attendees in-depth understanding of the scientific and technological issues at hand. The second part of the tutorial will focus on simulation, giving a state of art current simulation technology and discussing challenges for the development of sound simulation models.
The tutorial will use the SimGrid simulation framework as an exemplar since it implements sophisticated and validated simulation models. The last part of the tutorial will focus on an in-depth presentation of the different simulation approaches enabled by SimGrid, each with its specific range of applications and goals. SimGrid has been used to obtain results published in over 50 research articles and has thus emerged as one of the key tools for simulation in the area of parallel and large-scale distributed computing. Tutorial attendees will have the opportunity to gain some hands-on experience with SimGrid, by witnessing step-by-step development of small simulation projects. By the end of this tutorial, attendees should have a clear understanding of current technology and best practice for experimental parallel large-scale distributed computing research, and in particular on the use of simulation.
This tutorial is intended for attendees who are computer scientists with knowledge and interest in parallel and large-scale distributed computing. This tutorial should be extremely relevant for researchers interested in parallel and distributed algorithms (e.g. for applications scheduling, resource management, data management), which corresponds to a large subset of the researchers, attending PDCAT each year. Reasonable programming skills are expected, either in C or Java, so that attendees can follow example source codes.
Outline and schedule
Introduction (30 minutes)
- Needs and challenges for experimental large distributed computing research
- Comparison of current methodological approaches
- Real-world execution
- Hybrid approaches
Simulating parallel and large-scale distribution platforms and applications (50 minutes)
- State of the art relevant simulation tools
- Challenges for developing sound simulation models
- Case in point: simulation modules in the SimGrid framework and their validation
- Instantiation of simulated platforms
Using SimGrid (5 + 3 *25 minutes)
- Overview of SimGrid components and APIs
- The SimDAG interface: simulating the execution of DAG-structured applications
- An example simulation
- The MSG interface simulating asynchronous communicating process
- Goals of the interface
- Hand-on experience: step by step development of a small simulation project
- The GRAS interface: development of distributed applications for simultaneous simulated and real-world execution
- Goals of the interface
- Hands-on experience: step-by-step development of a small distributed application
Conclusion (20 Minutes)
- SimGrid success stories
- Perspective for the next 5 years
Arnaud Legrand is a junior CNRS researcher in the computer science laboratory of Grenoble (LIG). He works in the INRIA MESCAL project and his research focus on scheduling and performance evaluation of large-scale distributed systems. He has published over 50 research articles in peer-review journals and conferences. He obtained his M.S from the Ecole Nationale Superieure of Lyon, France in 2000, and his Ph.D. from the Ecole Nationale Superieure of Lyon, in 2003. Arnaud Legrand is one of the main developers of the SimGrid project since 1999.
Martin Quinson is an Associate Professor in the School of Computer Science and Applications of Lorraine at University of Nancy. His research interest are distributed, grid and internet computing. In particular his research emphasizes the development of distributed services over large-scale platforms, assessing the quality of distributed applications and the experimental evaluation of distributed algorithms. He has published over 15 research articles in peer-reviewed journals and conferences. He obtained his B.S from the Universite Jena Monet of Saint Etienne, France in 1999, his M.S. from the Ecole Nationale Superieure of Lyon, France 2000, and his Ph.D. form the Ecole Nationale Superieure of Lyon, France in 2003. He is a program committee member of the SIMUTools Conference since 2008 and of the CCGrid 2009 conference. Martin Quinson is one of the main developers of the SimGrid project since 2002.
Henri Casanova is an Associate Professor in the information and Computer Science Department at the University of Hawaii at Manoa. His research interest are in the area of parallel, distributed, grid, and internet computing. In particular, his research emphasizes the modeling and the simulation of distributed platforms and applications, as well as both the theoretical and practical aspects of scheduling problems. He has published over 70 research articles in peer-review journals and conferences. He obtained B.S. from the Ecole Nationale Superieure dElectronique, d’Electrotechnique, d’Infromatique et d’Hydraulique de Toulouse, France 1993, his M.S, from the Universite Paul Sabatier, Toulouse, France in 1994, and his Ph.D. from the University of Tennessee, Knoxville in 1998. Prior to joining the University of Hawaii, he was a Research Scientist at the San Diego Supercomputer Center and an Adjunct, Professor in Dept. of Computer Science and Engineering at the University of California, San Diego. Henri Casanova is the original designer and developer of SimGrid project.
Last modified: Monday, 02-Feb-2009 14:27:36 NZDT