Serpens for Kepler
Centre of Excellence for Global Systems Science
Global System Science (GSS) is an emerging research field with enormous potential. Its application extends to both private and public sector value, where industrial, economic, societal and scientific data can be harnessed and analysed to tackle global challenges that arise in energy, health, financial markets and urban development, to name a few.
To reach the impact of such value, several technical and organization challenges must be surpassed. First and foremost is the size: analysing these issues revolves around large scale, heterogeneous datasets, including synthetic populations that require extreme detail and accuracy. Processing such data presents great technical challenges in capacity and scalable performance to power its simulations, delivery and visualization.
HPC is the logical solution to this challenge. It represents an IT segment that is already actively addressing scientific fields, as well as positioning itself as the service line to meet Big Data demands towards a variety of user groups, whether being research communities, large enterprise customers or the public sector. IDC forecasts HPC to surpass 12€ billion by 2016, with at least 7% global growth over the next 3 years, although local European markets have been slower to meet the potential due to a large hit on the segment during the recent European economic crisis.1 HPC providers look to leverage the economies of scale: expanding their services and widening their customer base, and to do so, lower their operating costs. To maintain active growth, European HPC must become more approachable and affordable to the increasing demand from potential, yet currently unreachable, customers.
Scheduled finish date: 2018-09-30
Although the potential matching can be envisioned, current interaction between the GSS and HPC domains is hindered by organizational challenges and a gap in technical know-how and understanding. GSS scientists lack the ability to supply HPC-compatible code for deployment, and HPC providers lack the GSS-oriented environment and services to facilitate this entry point.
CoeGSS will bridge these technical, organizational and cultural gaps between HPC providers and GSS communities:
- On the technical layer, the CoE’s GSS-oriented HPC environments, toolsets, models and framework for application-specific global synthetic populations will enable HPC providers to service GSS customers, empower GSS researchers with the HPC resources needed, and provide innovative business opportunities for third-party SME developers to create added-value applications for a dynamic ecosystem of API-accessed datasets.
- On the organizational layer, the CoE’s single access point portal will converge stakeholders into an overlapping community of HPC providers’ GSS-oriented service and consultancy offerings, potential GSS researcher customers, and through them, analysis-consuming stakeholders in health, finance, energy, urban development, etc.
- Addressing the gap in knowledge, the CoE will focus on community building and training to educate HPC and GSS on their counterparts’ domains, increasing the economic and societal impact of both communities.
The above will position CoEGSS to ensure Europe’s continued HPC growth in a large scale, Big Data intensive research field, and likewise, enable the emerging GSS communities to deliver critical analysis on global problems in a variety of economic, health, environment and public sector areas.
On a high level view, the CoEGSS’s mission is to set up a collaboration platform and an infrastructure that can be applied for development of GSS applications that are capable of handling problems on the Big Data scale, leveraging the HPC storage and computation facilities. CoEGSS thus will serve as a unique platform, not only for Europe but also world-wide, which will allow development of GSS domain-specific applications in collaboration with the leading HPC experts. The platform will allow a trade-off between the traditional HPC technologies and requirements of the newly-emerging GSS data-centric development tasks.
The GSS centre provides preliminary models of the interaction between meteorological conditions, physical infrastructure and social networks. It visualizes the findings in animations, and due to its computing capabilities the visualizations need not hide the relevant uncertainties, but make them accessible for purposes of risk management. By convening experts on the medical aspects of heat exhaustion and related conditions as well as experts on mass communication, the centre helps to design efficient measures to help millions of people to deal with the risks of the situation. Thereby, the character of the GSS Centre as a distributed network enables it to avoid schematic “one size fits all” answers and to take into account the different contexts in which the emergency has to be addressed. Along these lines the GSS Centre can support policy makers to design and implement emergency responses and businesses – including media – to react in adequate ways to the short-term challenge.
- Establishing an advanced innovative HPC soft- and hardware infrastructure as a self-sustaining interface between stakeholders and GSS expertise, will result in the creation of the European GSS portal, which allows all stakeholders to get access to their needed level of information, services and contacts, whilst benefiting from the integration of state-of-the-art HPC technology and GSS processes.
- Enabling the processing/analysis of big heterogeneous GSS relevant datasets as base for GSS services, will result in toolsets for carrying out the processing, and in an information base, which will contain the specifications and documentation of the elements of the toolset.
- Provisioning of the needed technical HPC environment for GSS - we will produce a variety of established services and underlying technologies which cover the need of GSS applications, in particular a framework for generating application-specific global synthetic populations.
- In order to provision the necessary legal and economic frame for GSS on HPC we will develop a legal framework for the CoE, a suitable Business Model and Sustainability Plans, Pricing Schemes, and ROI Analyses.
- The set-up and maintenance of best possible support for GSS stakeholders, will result in accessible support services via the European GSS Portal.
CoEGSS presents to political and industrial stakeholders, and to the public at large, an HPC-based gateway to the results and methods of GSS. CoEGSS will use the established GSS methodology, consisting of the combination of computational tools and methods and structured dialogues involving both researchers and practitioners. It is essential to accept that there is no "one size fits all" synthetic population, but there are frameworks (methodologies and software environments) that allow the modellers to generate a specific synthetic population tailored to the specific needs of their simulation and to the specific data available.
The developed GSS platform will be directly integrated with the HPC systems made available to the scientific community in Poland. It will give an opportunity to find new groups of users to whom the GSS is directed. The new innovated platform and the knowledge concentrated in the CoE will be available at the time when Europe and US will face the transition to much more larger HPC installations and architectures different than the current. The overall aim is to provide a more user friendly environment, provide more functionality for a wide group of potential users and thanks to that to find extra HPC users interested in using the software platform. Furthermore PSNC will propose an innovative solution which will allow to use the GSS on top of HPC systems connected with cloud environment. This is an environment which may be of interests of industry, especially SMEs which would like to see new functionality but not necessarily used by very large systems. PSNC organizes annual end users conference on HPC, where the last achievements are presented. This conference will give us the chance to promote and exploit the knowledge to PSNC users. The knowledge of the CoEGSS represented by training courses will be combined with the trainings for users in PSNC. There is signed an agreement between all HPC centres in Poland, aiming to support the European PRACE initiative on HPC. PSNC as the main contractor of the agreement succeeded also to put the Polish PRACE activities on the RI roadmap in Poland. Thanks to the initiative PSNC is going to use the new exploitation channel to inform the academic community about the GSS functionality. The PIONIER association is the second communication channel to scientific community with the range to entire Poland (21 big academic communities in 21 biggest cities). The exploitation in commercial world will use communication channel from the Wielkopolska ICT Cluster (http://wklaster.pl/en/ ), converging 30 companies, mainly IT.