Researchers increasingly want to run computationally based scientific experiments on Distributed Computing Infrastructures (DCIs), seamlessly accessing large pools of computational resources and services. There are many production DCIs available, each requiring specific and detailed technical knowledge in order to manage the execution of applications. Unfortunately, DCIs tend to be highly incompatible with one another, and applications developed for one DCI are difficult to port to another. Workflow systems help shield the technical complexity of DCIs from researchers and are widely deployed as a virtualisation layer for many DCI’s. Workflows have become essential tools for expressing and sharing application expertise among scientific users, and in addition, providing a high-level execution model for infrastructure developers to optimise DCIs in support of research users. The use of workflow models significantly enhances a researcher’s ability to formalise and structure large-scale, high-throughput scientific experiments, accelerating the rate of scientific discovery. There exist many thousands of usable workflows but also many different workflow systems, and these are not interoperable. Learning the intricacies of any particular system in order to create and run workflows typically requires significant efforts from researchers, and it is unreasonable to expect researchers to learn new workflow systems every time in order to run workflows for a different system. Furthermore, researchers need seamless access to a range of DCIs to execute their workflows. Again, every DCI encountered presents daunting learning challenges. The newest generation of science gateways is able to meet these challenges and offers seamless integration of both multiple DCIs and multiple workflow systems.
Professor Terstyanszky’s lecture will expand on the benefits and challenges in the achievement of interoperability at workflow and DCI level, and how the challenges are currently being addressed within scientific gateway research, to support scientific users in the execution and management of large-scale DCI experiments, particularly in the cloud. The lecture will include a demonstration of the newest generation science gateway that allows researchers from different communities to easily and successfully run scientific experiments in the cloud, without the need for detailed knowledge relating to specific DCIs and workflows.
Professor Gabor Terstyanszky
Gabor Terstyanszky is a computer scientist and Professor of Distributed Computing in the School of Electronics and Computer Science at the University of Westminster, where he is the Director of the Centre for Parallel Computing. Before joining Westminster in 2000, he was a Principal Lecturer and Vice-Dean of the Faculty of Mechanical Engineering at the University of Miskolc in Hungary. In 2006, he was promoted to Reader at Westminster. His research in the field of distributed and parallel computing addresses the interoperability of large-scale distributed computing infrastructures (DCIs) and computational workflows. Working closely with diverse scientific communities, Professor Terstyanszky has contributed to breakthroughs in interoperability research that have enabled researchers in different fields to more effectively exploit DCIs in their scientific experiments, and to execute many types of computationally hungry and data-demanding applications such as computer modelling and simulation. Equally importantly, scientists may now more effectively collaborate with one another within their respective communities – by sharing complex workflows for example – massively enhancing the efficiency and throughput of their research. Researchers in fields as diverse as astrophysics, computational chemistry, heliophysics and life sciences are now able to easily share complex applications and migrate applications over multiple types of computing infrastructure, shortening the time it takes to produce research results, and producing many more research results than before. Professor Terstyanszky has supervised more than ten PhD students. He attracted in excess of £1.4 million in European and UK research funding to the University of Westminster. He has led or contributed to numerous collaborative research projects funded by European Commission research framework programmes and by UK research councils. He has secured several research grants at universities in Germany, Spain and the UK. He has published more than 100 papers in academic journals and in proceedings of European and international conferences. He is regularly invited to join programme committees of scientific conferences, and reviews scientific papers in academic journals on a regular basis.