Ad Emmen
AlmereGrid, The Netherlands
The IDGF Road Map for using Desktop Grids in eScience
Abstract:
The International Desktop Grid Federation has published a Desktop
Grids for eScience Road Map, earlier this year. It summarizes several
years of efforts to turn Desktop Grids into useful computational tools
for day to day large scale computational tasks. The Road Map provides
advise on all aspects, from technical to political oriented audiences.
An important part of the Road Map is dedicated to Green Desktop Grids.
While also legal asepcts of Desktop Grids are described in the Road
Map, it equally explains where Desktop Grids fit into the eScience and
Cloud landscape.
The presentation will concentrate on illustrating the implementation
of the Desktop Grid Roadmap in all kinds of situations, ranging from
small local Desktop Grids that allow for better use of a department's
resources, to large interconnected infrastructures, with several
Desktop Grids, and EGI type of Grids, bridged to allow seamless
integration in complex scientific work flows.
The International Desktop Grid federation organizes the Desktop Grid
Community ( desktopgridfederation.org ).
The work is supported by
the EDGI ( edgi-project.eu ) and DEGISCO projects
( degisco.eu ).
Pawel Gepner
Intel Technology Poland
The Datacenter of the Future – the road to Exascale
Abstract:
In 2002 there was a supercomputer that could hit 3.7 TFLOPS, that computer
took 25 racks, 512 servers and used 128 kW. By 2007 that same 3.7 TFLOPS
could be had by just a single rack's worth of computers. Today single rack
does more than 10 TFLOPS and consumes only 30kW so clearly Moore's law is
alive and doing well in the datacenter, but what does this mean to our HPC
challenges?
Today as we have commoditized the server, we have un-commoditized the data
center, but this is NOT a ''crisis'', rather it is an opportunity to
differentiate. Those who apply the right planning and engineering in the
exascale data center will prosper, those who do not will suffer.
Jamie Shiers
CERN, Switzerland
WLCG – the First Two Years of LHC Data Taking, Processing and Analysis
Abstract:
This talk will focus on the achievements of WLCG in the light of the first two
years of data taking at the LHC. Starting from Ian Foster's three-point
checklist for grid computing, it will analyze how well we have satisfied these
criteria and how grid computing weighs up with respect to alternative models
(clouds, centralized clusters and/or supercomputers). It will also analyze grid
use cases and motivate the need for retaining many but not all elements of the
current model for the foreseeable future. Finally, it will discuss future
directions not only for the HEP and WLCG communities but also in the wider
context of Europe's digital agenda. The talk will be given from a personal
viewpoint emphasizing the role of individuals and the benefits to science and
society that go way beyond the direct goal of providing a distributed computing
and storage solution to meet the needs of the LHC experiments.
Rod Hose
University of Sheffield, UK
Virtual Physiological Human Research and Clinical Simulation Workflows
Enabled by the VPH-Share Infostructure
Abstract:
VPH-Share is building an infrastructure for the sharing of tools, models and
data between members of the Virtual Physiological Human research community.
To provide concrete instantiations of the benefit of a coherent
infrastructure, the first target is to support and to disseminate four
specific workflows from existing projects (either completed or well-advanced
before the commencement of VPH-Share). This talk will include a discussion
of the state-of-the-art in development of VPH simulation workflows,
including the challenges of model personalisation and of clinical
translation. Currently our ability to produce geometrically faithful
representations of the individual patient anatomy is somewhat stronger than
our ability to represent individual physiology. One of the challenges is to
define the appropriate simulation envelope for the individual. The talk will
explore the requirement for establishment of an appropriate infrastructure
and architecture for meeting the needs of the VPH community.
Rob Meijer
University of Amsterdam and TNO, The Netherlands
Achievements of the UrbanFlod Project
Abstract:
It is scale that matters for the FP7 UrbanFlood (www.urbanflood.eu) early
warning system. It has been designed to monitor thousands of dikes and dams,
and one can use it also for a single one. It can handle hundreds of
thousands of sensors, and one can also use it for a measurement in your
backyard. Crucially, data driven modeling is deployed as there is no other
way to make sense of data from remote dikes and dams from which any details
are lacking. The keynote will reflect upon the UrbanFlood system as an
operating system for cyber physical infrastructures and details modern
computer science concepts that lie at its hart. This will be illustrated by
detailing and visualizing the applied concepts for cloud and network
architectures and AI.
Alfons Hoekstra
University of Amsterdam, The Netherlands
Distributed Multiscale Computing, the MAPPER project
Abstract:
Today scientists and engineers are commonly faced with the challenge of modelling,
predicting and controlling multiscale systems which cross scientific
disciplines and where several processes acting at different scales coexist and interact.
Such multidisciplinary multiscale models, when simulated in three dimensions,
require large scale or even extreme scale computing capabilities. The MAPPER project
develops computational strategies, software and services for distributed multiscale
simulations across disciplines, exploiting existing and evolving European
e-infrastructure. Driven by seven challenging applications from five
representative scientific domains (fusion, clinical decision making, systems biology, nano
science, engineering), MAPPER deploys a computational science environment for
distributed multiscale computing on and across European e-infrastructures. By taking
advantage of existing software and services, as delivered by EU and national projects,
MAPPER will result in high quality components for today's e-infrastructures. We develop
tools, software and services that permit loosely and tightly coupled multiscale
computing in a user friendly and transparent way. We integrate our applications into the
MAPPER environment, and demonstrate their enhanced capabilities. MAPPER integrates
heterogeneous infrastructures for programming and execution of multiscale
simulations. We reuse as much of the existing infrastructural and software
solutions as possible. The MAPPER solutions is developed on top of existing
e-infrastructures without the necessity to modify already deployed components. The
functionality to be delivered is realized as extensions to existing e-infra-structures. The
integration is done using well defined APIs and standard based interfaces, thus reducing
potential im-pact of changes on middle-ware level components.
Stephen Brewer
European Grid Initiative (EGI)
The impact of Virtual Research Communities on EGI's services and infrastructure
Abstract:
The Virtual Research Community model was established a year and a half ago at the time of
the creation of the European Grid Infrastructure as a federation of resource providers
across Europe and the start of the associated EU-funded project EGI-InSPIRE. Virtual
Research Communities or VRCs were envisaged as flexible organisations of scientists and
researchers who had a common area of academic interest for which having a single voice to
communicate their needs would benefit their research in terms of e-Infrastructure. The
mechanism for this communication is the User Community Board or UCB which meets monthly to
review the changing landscape of e-Infrastructure-based research in Europe. As a formal
channel into EGI, the UCB is empowered to guide the evolution of the infrastructure and to
advise on current activities especially with respect to support and delivery. Of course
communication operates in two directions and the community representatives are also expected
to act as messengers back to their own communities with insight into the operations and
plans of EGI and EGI-InSPIRE.
In the past the user communities were treated as more or less equally large blocks of
similar users. Whilst this was applicable under the large project framework which funded
support for grid research it is less applicable today. We live in the world of the long tail
and infinite customisation. People, especially researchers, expect to be be fully in charge
of their destiny. For this reason the services that EGI (and EGI-InSPIRE) offer to users and
communities are increasingly being designed to be scalable and flexible so that they are
equally applicable for the traditional Heavy User Communities such as High Energy Physics
and Life Sciences but also for smaller communities such as structural biology in the form of
WeNMR for example. Having established the EGI Training Marketplace, Applications Database
and the Requirements Tracker during the first year, we then went on to develop web gadgets
(also known as widgets) which can be embedded into your own or your group’s web site or
portal to provide a customised interface to these support-related resources.
As more research communities come on board as VRCs - both large and small, established and
emerging - so more flexibility and customisation will be built into the delivery of the
infrastructure. As the architecture of the infrastructure, in the widest sense of the word,
moves towards a multi-tiered model, so this personalisation will be come more powerful and
have a greater impact in terms of science and research across Europe and beyond. EGI is
listening and is evolving.
Krzysztof Gozdz
Hewlett-Packard Polska
The Cloud. Implemented
Abstract:
Clouds seem to be the next natural step in searching for maximal efficiency of computing systems, after SMP architectures,
clusters and grids. As they were treated as an idea for a few years, more and more persons are now looking for ready to run implementations.
HP, the company who has got the word ''cloud'' listed in the company strategy, offers various approaches to the solution,
from lending the infrastructure or applications, through migration of traditional data center models to the clouds, to building
modern converged infrastructure to be served in the cloud model. The presentation will show HP approach to the paradigm,
focus on the areas interesting in HPC and list a few real case studies, where the clouds are implemented.
|