|
The project Data Grid Explorer aims to implement a large scale emulation tool for the researcher communities of distributed operating systems, networks, and users of Grid systems. This large scale emulator consists of a database of experimental conditions, a large cluster of 1000 PCs, and tools to control and analyse experiments.
IDRIS, a French national supercomputer centre, will host the platform. The system will be connected to the French high speed academic network Renater 3.
The project includes studies concerning the Explorer instrument itself and others that deploy the instrument. The studies are grouped into four areas:
- Infrastructure
- Emulation
- Network
- Applications
The infrastructure task will in particular define the system and hardware, establish access methods, and describe how resources are to be shared. This task will look after incorporating flexibility into measurement techniques.
The emulation task will create a tool between simulation, which is purely software based, and experimentation that is carried out on the infrastructure. It will provide a facility for a wide range of studies of distributed systems at all architectural levels: large scale communication configuration, such as the Internet, and distributed applications, such as the Web, P2P rings.
This task will define the hardware and software equipment required to emulate the behaviour of network components. It will also define how to configure the platform for the particular context of a study, and how to place the monitoring equipment. Finally, it will allow the set-up of realistic experimental conditions. The emulation task is central to the Grid Explorer and will be used by many others projects.
The network task concerns all projects related to the design, the development, the evaluation, and the experimentation on new architectures and communication protocols, aiming to improve communications on the Grid or on the Internet. These improvements concern performance, quality of service, availability, security, flexibility, and manageability.
The applications task will study the porting of applications to the Grid and P2P infrastructures. It will target a broad range of applications fields including bio-computing and databases. The main goal of this task is to anticipate the performance problems that would be encountered in large scale configurations. |