|
|
|
||||
Aims and scope |
|||||
|
The VL-e project is now reaching its last year and a very important phase where one can observe the progress of several successful research activities, and at the same must also think about the future, after the project is finished. The objectives of this half-day workshop are:
The program will consist of demos, short talks, and a panel. The demos will be presented informally during lunch, coffee/tea break and drinks. The five short talks will highlight some of the demos or new research topics in VL-e. The panel discussion will attempt to raise key scientific issues for future e-Science research. Participation is open to VL-e partners and associated projects. Participation is free of charge, but registration is required. Please send an e-mail with your name and affiliation to Jacqueline van
der Velde (J.H.vanderVelde *at* uva.nl). |
|||||
Program |
|||||
|
|
|||||
|
List of
confirmed Speakers Scott Marshall (Informatics Institute, Tristan Glatard (Bioinformatics Laboratory, Academic MedicalCenter of the, Frank Seinstra (Department of Computer Science, VU University) Guido van Noordende
(Informatics Institute, Jan Just Keijser (Nikhef) List of
confirmed demonstrations per VL-e Sub-program SP1.3: Medical use case, Mirage, Compiled Matlab jobs on the grid SP1.4: Bird tracking system + ECOgrid SP1.5: A Problem Solving
Environment for Micro-array analysis SP1.6: KnowEx SP2.2: AIDA tools SP2.5: WS-VLAM, WorkflowBus, FRIPS SP3.1: GAT, IBIS, KOALA SP4: Resource Browser (VBrowser) and plugins (AIDA, WS-VLAM) Workshop program Lunch +Demo: (Neuwton room ) Session 1:
(Turing Zaal,
Room Z11 )
AbstractThe VL-e Medical
subprogram has developed a software platform to enable end-users to perform
image analysis experiments on the VL-e PoC
infrastructure. In this talk an overview of the architecture will be
presented, highlighting the recently developed services to execute workflows
on the grid and to access the gLite middleware for
job monitoring and data management. Usage statistics will illustrate the
regular adoption of this platform at the AMC by users from vlemed and lsgrid VOs..
Abstract As steadily
more organizations build data and knowledge repositories, there is a growing
interest in information extraction and knowledge capture technologies.
Researchers are developing components for knowledge management that
will enhance the ability to manage resources and perform computational
experiments. In the VL-e subprogram Adaptive Information Disclosure (AID),
researchers have combined information retrieval, machine learning, and
Semantic Web technologies to create the basis for a knowledge management
system in the AID Application (AIDA) Toolkit. Applications of AIDA components
are now available as Taverna workflows, a web
interface, and a plugin to the VBrowser.
Several of these applications to use cases will be described and
demonstrated, including use cases that demonstrate resource management and
knowledge discovery in biology and food informatics..
Abstract The Ibis
software platform drastically simplifies the process of programming and
deploying high-performance and distributed grid applications. The talk and
the demo both will demonstrate the power of the Ibis system by way of an
example from the research domain of automatic video content analysis. See
also http://www.cs.vu.nl/ibis/.
Abstract The VL-e PoC
infrastructure has been upgraded over the last year both at SARA and at Nikhef.
On the hardware side the compute clusters and storage facilities have been
upgraded, while a new version of the VL-e PoC
distribution software has also been released. This talk will focus on the
details of these upgrades and on some new VL-e applications which are now
running on the VL-e PoC infrastructure. Break +Demo: (Neuwton room ) 15:00-16:00 Session 2:
(Turing Zaal,
Room Z11 ) 16:00-17:30
Drinks + Demo: (Neuwton room ) 17:30-18:30 Demonstrations shown during lunch, coffee break and drinks
(SP1.3, Keith Cover) Description While the computational power to process large data sets in medical imaging hundreds of times faster is available through the GRID, the complexity of accessing this computational power has been a major barrier for many researchers in medical imaging. The Mirage GRID job distribution application was developed specifically to allow easy access to the computational power available on the GRID for researchers who want to learn as little as possible about the GRID. The main information the user needs to supply is (1) a Linux shell script and Linux executables that will be used to process each data set, (2) a list of the data sets to be processed and (3) the name of the GRID cluster or group of clusters to be used.
(SP1.3, Tristan Glatard,
Martin Stam) Description We will show live demos of the software tools that are used by medical imaging scientists autonomously at the AMC to perform large scale experiments on the VL-e PoC infrastructure. They include services for workflow execution on the grid and to access to the gLite middleware for job monitoring and data management.
(SP1.3, Matthan Caan) Description One of the experiments currently running in the Virtual Lab on Medical Imaging, uses compiled Matlab jobs, thereby overcoming the Matlab licensing problem. We will present the workflow, and a use case in reconstructing MRI-images of the human brain.
(SP1.4, Michael Kemp
(IBED) and Tijs de Kler (SARA) Description Virtual lab for Bird Migration Modelling We will show the Virtual Lab for Bird
Migration Modelling. In an international ESA project (with VLe partners IBED and SARA), we have (almost) real-time
access to military and meteorological radars and to the weather forecast. We
use these information sources to build models that predict the space-time
densities of migrating birds. Observations and model results are made
available to the operations centre of the air force. There this information
is used to change schedules of training to avoid bird strikes.
(SP1.4, Lourens Veen
(IBED)) Description The EcoGRID project
aims to make available species observation records from a wide range of
organisations in The Netherlands, and to process this data into scientific
knowledge to be used both by scientists and by non-scientific organisations
who need to comply to environmental legislation. We will demonstrate a portal
for presenting observation data to municipal ecologists, jointly developed by
the
(SP1.5, Wim de Leeuw and
Han Rauwerda) Description Micro-array experimentation requires a wide range of techniques and methodologies to pre-process, validate and analyze the data such as sequence alignment algorithms, statistical modeling, visualization, and set analysis methods. In the past years micro-array experimentation has matured and so has the analysis methodology. We developed a comprehensive system for microarray analysis centered at experimental design. Wherever needed, we have devised new tools and wherever possible we used de-facto standard methodology. Some methods were re-devised in order to make them suitable for grid computing. In this demonstration we will show the overall schema of the MA-PSE and we will demonstrate some parts of it.
(SP1.6, Ron Heeren) Description TO BE ANNOUNCED
(SP2.2, Scott Marshall, Marco Roos, Edgar Meij) Description The web service components of the AIDA Toolkit have been assembled into applications on three different platforms: Taverna workflows, a web interface, and the VBrowser (a java application with an AIDA plugin). The AIDA service-oriented architecture enables easier integration and more flexible development than monolithic applications typically allow. We will demonstrate a workflow for knowledge extraction and capture, a workflow that creates a Lucene index, a web interface that combines terminologies, ontologies, and document retrieval, and the AIDA plugin to the VBrowser.
(SP2.5, Vladimir Korkhov) Description The aim of the system is to provide and support coordinated execution of distributed Grid-enabled software components combined in a workflow. This system combines the ability to take advantage of the underlying Grid infrastructure and a flexible high-level rapid prototyping environment. The WS-VLAM workflow management system is developed following the OGSA/WSRF standards (http://www.science.uva.nl/~gvlam/wsvlam). The Demo will show how to design an application workflow, and run it on the Grid enabled computing resources. It will also show the new features of the workflow systems namely: workflow farming, hierarchical workflows, web service harvesting, semantic annotations of workflow components, and monitoring long lasting experiments.
(SP2.5, Zhiming Zhao) Description The demo will show VL-WFBus a FIPA compliant agent framework. The VL-WFBus provides interface to wrap legacy workflow engines, and coordination mechanisms to control the execution of the workflows. VL-WFBus has both a command line interface and Graphical user interface. The GUI of VL-WFBus is also available as a plug-in of the VBrowser. The coordination is developed based on language called Reo developed at CWI.
(SP2.5, Adianto Wibisono) Description FRIPS provides environment for scientists to manage Parameter Sweep Applications (PSA), change parameter space, set new policy for execution at runtime, and port legacy applications. This demo will show how this framework has been used for two application use case namely: Distributed Light Scattering (ADDA) and Functional Magnetic Resonance Imaging (fMRI).
(SP2.5, Sung Moring) Description The demo shows a software platform which enable to run Financial Trading and Research Algorithms. It will present aspects as real-time distributed event data caching for high speed financial data modeling, algorithm models distributable processed managed by a developed job services and how remotely users can control the parameters of the processed algorithms.
(Piter T. de Boer ) Description The Virtual
Resource Browser (VBrowser) started in 2006 as a
simple file browser for the Storage
Resource Broker (SRB) at SARA. Soon
after the first protoytype more grid file protocols
were added and the VBrowser grew from a simple file browser to a multi
client framework including other (non file) protocols which enabled browsing of 'virtual' resources and (grid)
services like they are part of the user's desktop. By creating an extensible framework the VBrowser
is now an integrating platform which
combines different application groups within the VL-e Project.
(SP3.1, Niels Drost, Roelof Kemp) Description In many emerging scientific and industrial applications the use of parallel and distributed computers at a very large scale is essential to meet the enormous, often petascale, computational demands. Ibis allows easy programming and transparent deployment of large-scale parallel and distributed applications, even for dynamic, faulty, and heterogeneous environments. In this demo we show the flexibility and power of Ibis using the Ibis Deploy system. Ibis Deploy allows the easy deployment of applications on grid systems. We will show several applications, including an object recognition system, deployed from a laptop to the grid. All required software is contained on a memory stick, and deployed on the grid transparently by Ibis Deploy.
(SP3.1, Alexandru Iosup, Hashim Mohemed, and Ozan Sonmez) Description The vast majority of jobs executed in current production grids is sequential. Many of these jobs are submitted in the form of Bags-of-Tasks (BoTs). We have extended the KOALA grid scheduler, which has been operational for three years on the DAS, with Cycle Scavenging (CS) support in order to have BoTs be able to put any remaining idle cycles in any clusters in a grid to effective use. This demo shows the use of the Grenchmark job submission and analysis tool and of the KOALA CS support, with as example application the Eternity II puzzle. |
|
||||
|
|
|||||
|
Date and location of the workshop |
|||||
|
· October 29, 2008. |
· Science Park Amsterdam. CWI - Turing Zaal Kruislaan 413 |
||||
|
Programme committee |
|||||
|
· Henri Bal (member of the VL e directorate, VU) · Cees de Laat
(Informatics Institute, UvA) · Silvia Olabarriaga
(Bioinformatics Laboratory, AMC, UvA) · Adam Belloum (Informatics Institute, UvA)
|
|
||||
|
Organisers |
|||||
|
Silvia Olabarriaga, Adam
Belloum & Jacqueline van der Velde email: S.D.Olabarriaga *at* amc.uva.nl, A.S.Z.Belloum *at* uva.nl, J.H.vanderVelde *at* uva.nl |
|||||
|
Informatics Institute, |
|||||
|
|
|
|
|||
