Project description

The Virtual Laboratory environment provides a framework for groups of scientists, engineers and scientific organizations that interact and cooperate with each other towards the achievement of a common experiment. Such an experimental environment enables researchers, at different locations, to work in an interactive way, as in any laboratory, i.e. the scientists are able to create and conduct the experiments in the same natural and efficient way as if they were in their laboratory.

One of the most important characteristics of the experimental domains is the manipulation of large data sets produced by the experiment devices. To be able to handle the resulting experiment data sets, three main requirements are supported within the VL architecture:

Proper management of large data sets: i.e. storage, handling, integration, and retrieval of large data sets. For example, in such a scientific environment, the size of data sets can range from few megabytes, in the case of the data sets generated by the Micro-array experiments, to tens of gigabytes, in the case of the data sets produced by the FTIR imaging micro-spectrometer.

Information sharing and exchange for collaboration activities: scientists are able to share both the devices used to perform the experiments and the data sets generated by those experiments. They must be able also to look at these data sets and compare them to the ones from previous experiments or other public databases, in order to find similarities and patterns.

Distributed resource management: the management of the resources must be properly considered in order to meet the high performance and massive computation and storage requirements.

Figure 1. Virtual Laboratory reference architecture

The Virtual Laboratory architecture, has incorporated these and other functional requirements through the design of different system components. In particular, the VL architecture consists of three main architecture components:

1. The application environment contains the scientific application domains considered in the VL (e.g. MACS application case, DNA Micro-array application case, and others), including certain specific domain functionalities.

2. The VL middleware enables the VL users to access low level distributed computing resources. The VL middleware provides: the VL user interface that enables the scientists to define and execute the experiments; the Abstract Machine (AM) that is the intermediate layer between the Grid infrastructure and the VL users; and three main functional components:

• The VIMCO component provides the functionalities to store and retrieve both the large data sets and the data analysis results, the advance functionalities for intelligent information integration and the facilities for information sharing based in a federated approach.
• The ComCol component provides the appropriate mechanisms for the data and process handling based on the Grid technology.
• The ViSE component offers a generic Virtual Simulation and Exploration environment where 3D visualization techniques are provided to analyze large data sets.
• The functionality provided by each one of these components is integrated through the VL integration architecture.

The Distributed computing environment provides the network platform that enables efficient usage of the computing and communication resources. Nowadays, a Gigabit Ethernet connection is being used. In the near future, it will be extended to a Wide­Area environment using GigaPort network based on the Surfnet5 backbone, which will result in a speed of 80 gigabits per second and a client connection capacity of 20 Gigabits per second. The Grid infrastructure provides the platform to manage data, resources, and process in distributed collaborative environments, such as the VL scientific applications. The Globus toolkit offers a set of tools to manage the resources in Data-Grid systems.

Contact person

Project Leader: L.O. Hertzberger (UvA)

H. Afsarmanesh (UvA)

P.M.A. Sloot (UvA)

J.F.J. van den. Brand (VU, NIKHEF)