LinkSCEEM-2
| The LinkSCEEM-2 Project |
 |
 |
 |
|
The entire Section 1 of the LinkSCEEM-2 Proposal in response to INFRA-2010-1.2.3: Virtual Research Communities can be found here. Concepts and Objectives Computation-based science and technology play a vital role in a broad range of research, engineering, education and commercial endeavours. As noted by the Partnership for Advanced Computing in Europe (PRACE): “[s]upercomputers are indispensable tools for solving the most challenging and complex scientific and technological problems through simulations.” The estimated world capacity as reported by the June 2009 TOP500 list is about 34 Pflop/s. Only two countries in the Eastern Mediterranean area have computers that appear in this list, namely Israel and South Arabia with a combined 0.6% of the total capacity. Since simulation has come to be regarded as the third pillar of science together with theory and experiment one may take it as another measure of how advanced scientific research is, and unfortunately according to this measure the region is rated well below Europe and the US. The lack of resources is one element that hinders the development of user communities. Others, and perhaps even more serious ones, are the lack of expertise and the limited integration of the scientific communities, so that the build-up of HPC infrastructure must be paralleled by the development of adequately trained human resources, both for technical and scientific personnel. The integration of the scientific communities must parallel that of the resources, and be supported by networking activities, and accompanied by a general raising of the awareness of the importance of HPC and computational science.
Indeed countries in the region are becoming aware of the importance of High Performance Computing (HPC) and the need to invest in HPC infrastructure and education. A number of complex problems of high societal and/or economic impact such as, for example, providing reliable predictions for climate changes in the next decades can only be addressed by large-scale computation that requires supercomputers. Of crucial value in science and engineering is the ability to describe phenomena at various length and time scales. In many problems, the relevant scales that have to be considered differ by many orders of magnitude. Such applications, that are abundant in, for example, material science, biochemistry, weather prediction and life sciences, require appropriate numerical methods and access to supercomputers. As already mentioned, simulation has in fact become as important as theory and experiment and in fields like material sciences “one effectively performs 'computer experiments' to observe structures, properties and processes and to conduct measurements that cannot be readily undertaken in the laboratory”. The ability of scientists in the Eastern Mediterranean region to produce new knowledge and contribute to the world research output in these important fields relies on the build up of state-of-the-art know-how in simulation science and HPC infrastructure in the region.
The economic impact of computers in a very large range of activities, from banking to travelling, has been increasing significantly in the last two decades, so that besides being indispensable to science and technology, high performance computing is also becoming crucial in finance and the humanities. Financial risk management is a good example of a sector of the economy that requires significant computational resources and techniques. This is an additional compelling reason why HPC infrastructure in the region is urgently needed.
Educational approaches are also evolving with computer advances and increased connectivity. Virtual schools and workshops are being developed to integrate human resources and reach out to small universities and research communities. Therefore if the region is to educate the next generation of scientists and engineers and reduce the brain drain it is experiencing, e-infrastructure development is a necessity. In agreement with the rationale and objectives of the European Neighbourhood Policy, and of the Union for the Mediterranean, Europe and the European Research Area only benefit from the development of communities of highly trained scientists and engineers at its periphery, drawing on the human capital of over 100 million people.
The FP7 Support Action LinkSCEEM has initiated the development of a network of users in various fields who will benefit from the development and integration of HPC facilities. It has carried out surveys aimed at the assessment of regional needs and demands, and organised a number of user meetings for direct contacts with the research communities. A clear need for enhanced e-resources, with associated user support, was demonstrated, while training and educational programs were earmarked as important activities for many scientists in order to be able to use Tera- and Peta- scale computing.
The present project is a follow-up and implementation phase of LinkSCEEM. It will follow a three fold approach:
Regional climate change simulations as well as the digitisation and 3-dimensional visualisation of cultural heritage are example areas where this project will bring state-of-the-art expertise and where new computer architectures can provide a significant advantage with high societal and economic value to the region. SESAME, the first synchrotron light source facility in the Middle East is developing a very broad user community throughout the region. It will not only bring to the region the benefits of synchrotron light applications in many scientific fields, but also contribute a culture of scientific cooperation that could have enormous impact on the regional science and technology scene. The SESAME council has decided to use the computational resources of CaSToRC to process the data produced by the SESAME beamlines, and the present project will make a crucial contribution to that objective, through the provision of resources, user support and training.
Three regional HPC facilities will contribute resources and services in this project. This proposal aims at providing funding for activities that go beyond the services that these facilities offer to their institutional and national users, and are aimed at engaging virtual research communities at the regional scale. It is worth noting that in all three cases the “hard” operational costs associated with the compute cycles that will be provided (e.g. for power) will be covered by national funding, while the proposal budget will cover notably the costs associated with enhancing the user support and peer review capacity of the 3 regional HPC centres. The Cyprus Institute (CyI), has launched a Computation-based Science and Technology Research Centre (CaSToRC) being developed in partnership with the National Centre for Supercomputing Applications of the University of Illinois. CaSToRC will include a high-performance computing facility, starting in the tens of Teraflops and developing to the hundreds Teraflops scale, with an associated research and educational programme devoted to computational science. At the final stage of development CaSToRC will be comprised of three major divisions: a high performance computing centre, a computation-based science and technology research division and a cyber-infrastructure (e-science) division. The Bibliotheca Alexandrina (BA) has recently installed a SUN cluster of peak performance exceeding ten Tflop/s, and also has an extensive archival system. In addition, its visualization infrastructure, incorporating an active stereo immersive environment, is one of the most advanced in the region, and the BA conducts joint activities and partnerships in visualization projects with Universities in Egypt, the region and internationally. Besides CaSToRC and BA the HPC facilities at NARSS, currently consisting of a one-rack Blue Gene/L, are expected to be upgraded to meet the needs of the larger user communities that will be developed. Its resources also include a visualisation lab, a satellite station and facilities for remote sensing and satellite image processing.Additional sites are expected to be developed and integrated into the project, such as the computer resources that are being planned for on site processing of SESAME data, as well as the computer resources of several Israeli institutions.
Through the activities undertaken in LinkSCEEM a significant network of scientists has been established in the area. The present project will further develop and expand the scientific links among scientists in the region as well as the collaboration of the Eastern Mediterranean research teams with the corresponding ones in Europe and the USA. In particular, this project will promote collaboration between the ESRF and SESAME facilities in the area of Synchrotron radiation (SR) research, between Cyprus/Egypt/Israel and the Mainz Max Planck Institute in Climate studies and among NCSA, Bibliotheca Alexandrina, CaSToRC, NARSS and their associated teams in visualization and cultural heritage. All HPC facilities and research institutions of the consortium are currently engaged in user support, training and research activities and therefore this project will provide additional resources in terms of expertise and human resources, in order to enlarge and integrate the user communities in the region. It will promote further collaboration among the partners, integrate computational resources and optimize their usage. In the case of visualization, it will made accessible for the first time advanced and collaborative infrastructure for scientists in the region to pursue activities in digital cultural heritage, presentation of climate data, and synchrotron based research. As pointed out, this project will focus on the research communities in three fields that a) are relevant to the region and at the same time are of global importance, and b) in which significant activity already exists in the region (notably at some of the partner institutions), which will guarantee that the proposed activities have maximum impact: Climate Modelling: There are now robust predictions, notably from the Intergovernmental Panel on Climate Change (IPCC) showing that the increase in temperature due to global warming will be particularly severe in the Eastern Mediterranean region, resulting notably in the worsening of water scarcity. These issues are currently being further explored by the CyI Climate Group, notably within the C8 project funded by an ERC Advanced Grant2. Obviously, such climate changes will have enormous societal and economic consequences, and policy makers need to be informed in order to take appropriate measures. One of the key issues concerns more accurate climate change predictions, relying on higher resolution models taking into account the chemistry of the atmosphere. There is a well-established group at the Cyprus Institute, working in close collaboration with the Max Planck Institute on Atmospheric Chemistry (MPIC) in Mainz, and with links to research groups in Israel, Egypt, Greece and Turkey. This, together with the climate simulations also carried out at NARSS, constitutes a very favourable basis for engaging a broad regional user community that can highly benefit from the resources at CaSToRC, BA and NARSS, and the expertise provided by the consortium partners, notably MPIC. Digital Cultural Heritage: The preservation and showcasing of cultural heritage in digital format constitutes one of the priorities of the EU. The Eastern Mediterranean region possesses a rich cultural heritage. The history and culture of the area is among the oldest in the world. If writing is taken as a prerequisite for civilization, then the earliest "cradle" is Early Dynastic Egypt 3100 BC. The first signs of civilization in Cyprus traced in archaeological excavations and research date back 9,000 years to the 7th millennium BC. This rich cultural landscape involves hundreds of archaeological sites scattered throughout the region, representing various historical periods in the region‟s evolution. The EU's digital libraries initiative sets out to make all Europe‟s cultural heritage resources accessible to all, and preserve it for future generations. It is important for the region to develop a similar project creating a digital library that will function as a single, user-friendly access point for the collections. It is imperative that digitization of Cultural Heritage efforts in the region are accelerated in a coordinated manner to match similar activities in the EU countries. CaSToRC has been established as a research centre in the European landscape of Digital Libraries with a number of significant research projects3. The CyI is in the process of establishing Kypriana, which will work in collaboration and as a portal to Europeana. The Bibliotheca Alexandrina has a long-standing involvement in digitization of ancient text with an established infrastructure in digitization and visualization. The Digital Laboratory of the library has been very active in the management and curation of digital texts and objects. Synchrotron radiation research: Synchrotron radiation facilities like SESAME and ESRF enable scientific research in a broad variety of fields: atomic and molecular physics, biology, material science, spectroscopy, archaeology and imaging. They therefore help develop cross-disciplinary research leading to a large range of potential applications. For example, x-ray imaging can offer unprecedented resolution for several applications, including medical imaging, security screening and industrial non-destructive testing. Studies of novel materials can lead to the construction of new devices such as thin films with controllable electronic properties having potentially large impact on future applications in sensors and computing. The first measurements from SESAME are expected in 2012 and by then the software for the analysis and expertise on data management of large data sets should be in place. This project will contribute expertise and integration of computer resources for the analysis of SESAME data. Transfer of data from SESAME to the HPC facilities requires improvement in the connectivity. In this project we propose, as first step, a line to be established between Cyprus and Jordan, building on the available bandwidth and with an adequate ramp-up for the transfer of data from SESAME to CaSToRC. The user communities in the above fields will greatly benefit from the networking, resources and services proposed in this project. This is a conclusion that can be drawn from surveys of user needs in the various fields undertaken under the LinkSCEEM project. The survey results indicate that most scientists are using PC/Workstations or small clusters (over 80% of identifiable infrastructure) and most of them expressed the significant need for more computers resources but also, as already mentioned, for professional user support and training. Unlike in Western European countries, access to Teraflop/s computers is almost non-existent and training in the usage of multi-core architectures, with a few exceptions, is very scarce. Additionally, although interest in cultural heritage is present in most countries in the region, digitisation and visualization expertise and infrastructure are not available. The activities carried out by this project therefore reflect the interests and needs of the regional communities as follows: a) user support and training ranging from basic to advanced level will provide the necessary HPC skills for users of the cluster machines at CaSToRC and Bibliotheca Alexandrina, as well as for applications that can be ported on the Blue Gene/L machine at NARSS, b) cross-disciplinary activities addressing development of algorithms and programming tools for Tera-scale computing, as well as data management and visualization techniques will complement the training by providing more specialist support for the research activities in the thematic areas of this project and c) development of the user communities in the three thematic areas by supporting high quality service activities and international level research in Climate Studies, Digital Cultural Heritage and Synchrotron radiation source applications. These activities will prepare regional researchers in these areas to use the Teraflop/s machines effectively, and a subset of them to be ready to migrate to Peta-scale computing by the end of the project.
The infrastructure at CaSToRC, Bibliotheca Alexandrina and NARSS will provide for most users the first opportunity to use supercomputers instead of office workstation /PC or small clusters. The one-group one-cluster model is the starting point for many applications. However, in order to make use of capability computing and for optimal usage of the computational resources, access to common infrastructure and availability of user support for development and optimization of codes is indispensable. In addition, the project will contribute to the integration of compute resources with the data and visualization resources that are required for applications to have optimal scientific impact. Moreover this project will contribute to the sharing and coordination of methodologies and practices, catalysed by the presence of Tera-scale systems and by the transfer of high-level expertise and know-how from internationally reputed institutions. This in turn will create greater integration between the resources that are actually involved in the project and the local clusters operated by research groups in the region, which will favour the development in the region of a HPC eco-system, similar in spirit to that which PRACE is promoting in Europe, but at a more modest scale.
In summary the main objectives of this project are to interlink and coordinate regional computational, data and visualisation resources to form an integrated e-platform, as well as provide the associated training activities and user support, and engage regional communities through networking activities such as workshops, exchange of visitors, organisation of joint events, and outreach. This will contribute to the creation of an eco-system of HPC from the individual group cluster to the Teraflop scale machines provided by CaSToRC, Bibliotheca Alexandrina and NARSS, and promote the coordination of practices and methodologies. The specific goals are:
|