12-18 November 2005, Washington State Convention and Trade Centre, Seattle, USA
Stand Number 922
Titles and Abstracts
Prof. Malcolm Atkinson, University of Edinburgh
Exploiting Diverse Data Resources the OGSA-DAI way
The wealth of data sources available for research is growing rapidly in all disciplines. Adept researchers can and should exploit these resources to accelerate discovery, but there are many arduous challenges that stand in their way. OGSA-DAI is the de facto standard foundation for building data integration solutions. It is used by 1500 users and many major projects. In this talk we answer the question "What can it do for me?". We describe the latest middleware architecture and present planned developments. We propose a framework and vision for international collaboration in the construction of a richer data integration environment.
Prof. Malcolm Atkinson, University of Edinburgh
International Collaboration to Extend and Advance Grid Education
Major advances in the provision of e-Infrastructure and CyberInfrastructure are opening up new ways of working. Now we need to develop a wide understanding of how these opportunities may be exploited. The ICEAGE project will stimulate development of the necessary education across Europe and beyond. It will support educators developing new curricula in many disciplines and will encourage adoption of those curricula in university programmes. We present our first thoughts on how this may be done and seek dialogue and partners in developing this educational agenda. By joining this collaboration you can help steer the work and share the results. This will be worthwhile as it will catalyse the application of emerging e-Science methods for the benefit of society.
Prof. Peter Coveney, University College London
Simulated Pore Interactive Computing Environment (SPICE): Using Grid computing to understand DNA translocation across protein nanopores embedded in lipid membranes
SPICE aims to understand the vital process of translocation of biomolecules across protein pores by computing the free energy profile of the translocating biomolecule. Without significant advances at the algorithmic, computing and analysis levels, progress on problems of such size and complexity will remain beyond the scope of computational science for the foreseeable future. Grid computing provides the required new computing paradigm as well as facilitating the adoption of new algorithmic and analytical approaches. SPICE uses sophisticated grid infrastructure to couple distributed high performance simulations, visualization and instruments used in the analysis to the same framework. We describe how we utilize the resources of a federated trans-Atlantic Grid to enhance our understanding of the translocation phenomenon in ways that have not been possible until now.
Dr Dave Colling, Imperial College London
GridPP deployment and operations.
GridPP has 19 sites supplying about 3000 job slots to the Large Hadron Collider (LHC) Computing Grid (LCG). The deployed resources, which include 1PB of tape storage at the main (Tier-1) site, will increase significantly over the coming year. GridPP operations and deployment take place in collaboration with Enabling Grids for E-sciencE (EGEE) and the UK and Ireland Regional Operations Centre (ROC). We monitor day-to-day operational performance using Site Functional Tests that indicate whether or not critical site components are working. If the tests report failures these are followed up by the deployment team. To help ensure stable operations and discover problems early, GridPP is taking an active role in the LCG service challenges which are designed to stress test and develop the high-bandwidth infrastructure required by LCG. This talk will examine on the current status and plans for GridPP's deployment and operations, including reporting the latest results from LCG's Service Challenge 3
Dr Roger Jones, Lancaster University
GridPP applications and portals
The very large data volumes and processing needs of the Large Hadron Collider experiments will require extensive use of Grid technologies. Each of the four experiments (ATLAS, CMS, LHCb and ALICE) has devised computing frameworks and Grid applications are now being developed to allow these frameworks to work in a distributed environment. The UK has aimed to find commonalities between the needs of the various experiments, and to co-ordinate effort of shared benefit where possible, and this talk will consider two examples. Firstly, the GANGA project is developing common components for the two experiments ATLAS and LHCb. GANGA is providing the front end to the distributed analysis systems for the experiments, making Grid access transparent and allowing the user to configure their jobs in a sophisticated way. The second example is the GridPP Grid portal; it provides a toolkit based on web services and template implementations, allowing experiments with few resources to build a Grid interface quickly with little prior expertise.
Dr. Jonathan Essex, University of Southampton
The dynome: Exploring protein dynamics by terascale simulation
The way in which proteins move and flex is critical for their function. Unfortunately, the main experimental method for probing and determining protein structure, X-ray diffraction, provides comparatively little information on how the protein moves. Molecular dynamics computer simulations, on the other hand, are able to yield detailed information on protein dynamics that is very useful, particularly for rational drug design. As an exemplar of this paradigm, large-scale high-performance computers will be used to simulate the dynamics of a range of membrane-bound proteins. Studies of how specific mutations (for example, drug resistant mutations) affect the motions of HIV-Reverse Transcriptase will also be performed. These simulations will take place as part of the SuperComputing 2005 scientific meeting, to demonstrate the power and importance of integrating large computing resources. The computer software used to control the simulations, store their output, analyse and present the results, and select new simulations to be run, will be demonstrated. These studies are important not only for the computing community itself, but are also particularly timely because of recent development in protein X-ray structure determination. With the arrival of the Diamond Light Source in 2007, and the stated aims of the UK Structural Genomics Consortium (www.sgc.ox.ac.uk), the number of new protein structures available will increase significantly. This proposal will demonstrate how simulations may be directly linked to this experimental information, providing critical extra data on the protein dynamics.
Prof. David de Roure, University of Southampton
The CombeChem Semantic DataGrid
The CombeChem project has designed and deployed an e-Science infrastructure using a combination of Grid and Semantic Web technologies. This 'Semantic DataGrid' provides a rich record of scientific and experimental data and its provenance, and a platform for sophisticated scientific queries.
Carl Christensen, University of Oxford
ClimatePrediction.net: Predicting Climate Change Through Volunteer Computing
The climateprediction.net (CPDN) experiment distributes state-of-the-art climate models to the public to run on their home, school and work computers. This was done in the manner of the forerunner volunteer computing project SETI@home, and now uses the Berkeley Open Infrastructure for Network Computing (BOINC). The results will allow scientists to make the most complete probability-based forecast of 21st century climate attempted to date, and should allow decision makers to plan with greater confidence for possible consequences. Since the launch in September of 2003; 90,000 people have joined the experiment in over 130 countries worldwide. The first results of the experiment were published in the science journal Nature in January of 2005 (27 Jan 2005 issue), and show climate sensitivities from thousands of completed runs ranging from less than 2 K to more than 11 K degrees.
Dr. Jon Blower, Reading e-Science Centre.
Data streaming, workflow and firewall-friendly Grid Services with Styx
A common aim in e-Science and Grid computing is to be able to construct distributed applications by composing remote services together in a workflow. This requires that data be transferred across the Internet from service to service. Many systems (e.g. many that are based on Web Services) require not only that the data be encoded into XML (greatly increasing the data size) but also require the data to pass through the workflow engine. This is only generally acceptable if the datasets are small. We have developed the Styx Grid Service (SGS), a service type that allows data to be streamed directly from service to service, greatly increasing the efficiency of workflows. SGSs can interoperate with Web and Grid Services through the use of wrappers. The SGS architecture allows the monitoring of quantities such as the progress of each service, without requiring incoming ports to be open through the firewall. We shall also demonstrate that Styx Grid Services can also be use for collaborative visualization and computational steering.
Dr. Steven Newhouse, Open Middleware Infrastructure Institute
The OMII 2.0 Grid Distribution: What it can do for you!
The OMII 2.0 provides an easy to install, use and manage software distribution to enable inter-organisational collaborative computing. Based around an open-source web services hosting environment, this is the first release to include components from the OMII's managed programme. The presentation will identify the capabilities of the current distribution and how it will evolve to further meet the needs of the community.
Dr. Martin Dove, University of Cambridge
Integrating compute and data grids within the eMinerals collaboratory
The eMinerals project involves a collaboration between molecular simulation scientists, code developers and computer scientists. The aim of the project has been to develop a grid infrastructure that provides both computer and data grid facilities, and to provide support for tools to enable the collaboration to operate as a virtual organisation. We had two primary design requirements. The first is the need for the infrastructure to support heterogenous applications, and this has meant working with heterogenous compute architectures by choice. The second was for the infrastructure to be easy for the scientists to work with. In brief, the solution to these issues was to develop a compute grid infrastructure based on both Globus and Condor, specifically developing the user interface through Condor-G. To overcome problems associated with data transfer, we have built the data grid component using the San Diego Storage Resource Broker. We have developed a number of tools, including job submission scripts, a job submission and data management portal, and a web-services interface, based on these tools. To make progress, it has been essential to develop the means for regular collaboration between various members of the project. We have made considerable use of the desktop version of the Access Grid, and have developed a general-purpose multicast application-sharing tool.
Prof. B.M. Boghosian, Tufts University
VORTONICS: Grid Computing for the Simulation and Identification of Vortical Structures in Fluids
Vortex knots and links are the most evident coherent structures of fluid turbulence. Like the elementary particles of high-energy physics, there are a wide variety of such structures, more exotic ones are found at higher energies, and they have a characteristic lifetime after which they decay into other such structures. An improved understanding of the creation, behavior, and interaction of vortical structures would elucidate fundamental questions in theoretical fluid dynamics, such as the reconnection problem and the detachment of “hairpin” vortices from boundaries in the intermittency approach to turbulence. It holds the potential for improved understanding in applied fluid dynamics, including the dynamics of meteorological structures such as tornadoes and hurricanes. It may also lead to improvements in computational fluid dynamics, such as naturally adaptive vortex algorithms that avoid vortex tangles and requisite “vortex surgery.” High-resolution direct numerical simulation of the dynamics of vortical structures at high Reynolds number are, however, among the most intensive scientific computations attempted today. We describe the VORTONICS package for locating and tracking vortex cores, and addressing other fundamental problems of topological fluid dynamics. The package includes modules for generating, evolving, and identifying vortex cores, and routines for Fourier-resizing and remapping the computational lattice. It is parallelized using MPI, and geographically distributed using MPICH-G2. It also includes a methodology for computational steering, checkpointing, rewinding, and dynamic adjustment of resolution. We describe the geographical distribution of the tasks performed by this code, as well of the domain-decomposition of the computational lattice itself on the TeraGrid and other distributed HPC resources.
Prof. G.E. Karniadakis, Brown University
TeraGrid Cross-Site Simulations and Visualizations of the Human Arterial Tree
Cardiovascular disease accounts for almost fifty percent of deaths in the western world. The formation of arterial disease such as atherosclerotic plaques is strongly related to the blood flow patterns, and is observed to occur preferentially in regions of separated and recirculating flow such as vessel branches and bifurcations. In this talk we will perform for first time simulations of blood flow in the entire human arterial tree through detailed three-dimensional computations at a number of arterial bifurcations, coupled by the wave-like nature of pulse information traveling from the heart to arteries that is modeled by a reduced set of one-dimensional equations. We employ MPICH-G2 and conduct geographically-distributed coupled cross-site simulations at major TeraGrid sites in the US and high-end systems in the UK. Flow visualizations on all arteries simultaneously will also be demonstrated using the TeraGrid resources.
Prof. Peter Coveney & Dr. Gianni De Fabritiis, University College London
Coupled models on the Grid
Coupled models are set to become increasingly important in all aspects of science and engineering as a means to study complex systems in an integrative manner, particularly in biology and materials science. In computational terms, our aim is to embrace, within a single simulation, all necessary length and time scale processes to describe the system of interest. Such coupled, hybrid simulations typically communicate data between the component models of which they are comprised relatively infrequently, and so a Grid is expected to present an ideal architecture on which to run them. Computational steering is a powerful tool for interacting with simulations as they run, allowing the user to alter physical parameters, algorithmic parameters or even to migrate their simulation to another machine on the Grid. Hybrid models present new challenges in these respects. In this talk and demo, we describe a simple, flexible and extensible architecture for coupled models based on the RealityGrid computational steering system and discuss its implementation for the case of a two-component hybrid molecular-continuum fluid dynamics model.
Prof. Yike Guo, Imperial College London
Bridging the Macro and Micro: A Computing Intensive Earthquake Study Using Discovery Net
We present the development and use of a novel distributed geo-hazard modeling environment for the analysis and interpretation of large scale earthquake data sets. Our work demonstrates, for the first time, how earthquake-related surface deformation measured from satellite images using imageodesy algorithms is coupled with analysis and simulation using finite-element numerical models. Our work realises a real time distributed analytical environment where analysis and simulation are closely coupled; integrating high performance implementations of image mining components executing on dedicated Discovery Net servers at Imperial College London, UK and high performance implementations of finite-element models executing at specialised servers at the University of Oklahoma, USA. Novel scientific results produced using our data sets provide a valuable insight into earthquake analysis. In addition, our informatics work provides a novel high performance computing framework and methods for the application of complex knowledge discovery methods to understanding earthquake dynamics. Furthermore, the realisation of our distributed computing platform is based on the implementation of a set of open standards, making its results accessible over the Grid to the wider scientific community.
Dr Andrew Richards, NGS Coordinator
The United Kingdom National Grid Service: Building a Production Grid
The UK National Grid Service (NGS) is funded to provide a significant resource for the UK e-Science Community and is result of funding from the UK e-Science program. Over the last 1.5 years the NGS has grown into a production service and has incorporated several additional sites to provide a heterogeneous grid that operates around a common middleware and a common set of policies. Looking forwards the NGS aims to continue to attract researchers from diverse academic backgrounds and to strengthen the integration with other international grids such as TeraGrid and EGEE.
Keith Noddle, University of Leicester
AstroGrid: Providing Pervasive Access and Integration of Astronomical Data and Applications
Modern astronomy is on the brink of some really exciting new discoveries. This has been made possible by a new generation of extremely powerful instruments, both ground and space based, which are providing unparalleled views of the universe - and at all wavelengths of the spectrum. These instruments are gathering masses of data - up to terabytes before reduction. These data are the core of the new research being undertaken, but data holdings are growing faster with each new instrument and the scientific research is becoming ever more complex as astronomers seek to study the sky in as much detail and across as many spectra as they can.
Making effective use of all these data is the key challenge astronomy now faces and the answer is the Virtual Observatory; a worldwide collaboration between observatories, university, research institutions etc to provide consistent, usable access to their data and resources for all astronomers.
This talk will examine how AstroGrid provides a complete answer to this challenge.
Dr Tom Jackson, University of York
Cortex II: A parallel FPGA machine for data search and analysis applications.
The Cortex II parallel computer is a system specifically developed to provide a flexible platform for data intensive search and analysis applications. The system consists of an array of PRESENCE II processor nodes, each contains a large FPGA (6M Gate, Vertex II), a TI DSP and up to 4Gb of memory. The FPGA sits at the heart of the system, connecting the main bus (PCI based), Memory and DSP. The system is hosted in any machine with a PCI bus, our current host machine is a Sun enterprise server. The number of nodes is only limited by the available slots on the chosen host machine. The system may be used as a bare machine, where support is given to the system resources on the FPGA, DSP, bus and Memory. Cortex has been designed to support the AURA high performance pattern matching system. The technology allows rapid searching of large unstructured data sets. The main search engine supports Vector and Scalar numbers and characters as well as a graph search capability. Application adapters are provided for text searching, signal search, face recognition, logo matching and molecular searching. The system is being evaluated in many problems including searching signals in large databases of Rolls-Royce engine data, for 3D face Recognition in security and access control applications and for pharmaceutical problems including Mass Spec data search and Drug Discovery.
The Cortex II machine has been developed by the Department of Computer Science at the University of York, UK and is now being sold by Cybula Ltd.
Dr Eduardo Gonzales-Solares, University of Cambridge
Managing Communities in the Virtual Observatory: a global system for authenticationand authorisation for astronomy
The International Virtual Observatory (IVO) is a widespread virtual organization. Compared to production compute-grids such as EGEE, UK NGS and TeraGrid, the IVO is both more diverse, in its funding and membership, and more focused, in that it serves a single domain of science. This talk shows how access-control techniques from grid computing have been adapted to the IVO case. In particular, we see how autonomous communities are use to decentralize and streamline management of users, PKIs, user-groups, and proxy certificates.