Introduction

Operating system developments must be validated on real hardware. The behavior of a grid operating system depends on so many parameters such as the number of nodes, their heterogeneity (memory, cpu, devices), the structure of the Grid (small or large clusters), the interconnection network (structure, latency, bottlenecks), the dynamicity of the grid, the stability of the grid (node failures), the efficiency of grid services, service bottlenecks etc. that their is no possible way to evaluate it through simulation.

Grid operating systems such as XtreemOS must be validated on a realistic grid platform.

The grid testbed must provide a significant number of computation nodes for scalability evaluation. The testbed nodes must allow full reconfiguration of the software stack from the low level communication layers up to the grid services.

Grid'5000

The Grid’5000 platform (http://www.grid5000.fr)  is distributed on 9 sites in France, provides a significant number of computation nodes (>1000 dual nodes end 2006, expected 2500 dual/quad nodes end 2007) and allows full reconfiguration of the node software from the operating system kernel to the grid middleware. Grid’5000 users can reserve nodes on Grid’5000 sites, define new operating system environments (kernel and distribution) and deploy these environments on the nodes.

Grid’5000 is the initial testbed for the XtreemOS project.

Extension of the testbed

The integration of DAS 3 and CNGrid to the initial testbed is planned for the second half of the project.

The DAS-3 system (http://www.asci.tudelft.nl/) will have five clusters with about 200 dual-CPU PCs in total,  which are integrated into a single large-scale distributed system using the multi-colour optical networking capabilities from GigaPort. Four clusters will be located at the following ASCI universities: the Vrije Universiteit (VUA), the University of Amsterdam, Delft University of Technology and Leiden University.

The China National Grid (CNGrid) (http://www.cngrid.org/) serves as a nation-scale testbed for grid technology research and development in China, which is sponsored by China Ministry of Science and Technology. By August 2005, the CNGrid project has built an experimental environment, consisting of eight grid nodes (computing and data centers) spanning six cities in China. The total computing capability exceeds 20 Tflop/s, provided by domestic HPC systems such as Dawning 4000A (10 Tflop/s) and Lenovo 6800 (5 Tflop/s), as well as HPC systems from multinational vendors.

Other platforms from XtreemOS partners can also be integrated in the testbed.

Publications

Project deliverables

• First version of XtreemOS testbed user manual (D4.3.1) - December 2006
• Study of XtreemOS test bed extension (D4.3.2) - December 2007