Vier Wochen lang besuchte Vasiliy Chitanov von der bulgarischen Akademie der Wissenschaften die RISC Software GmbH um mehr über den Einsatz von Grid Computing zu lernen.

“Als ich 2010 das erste Mal in Hagenberg war und an einer Summer School für symbolisches Rechnen teilnahm, war ich sofort von der Infrastruktur und der netten Arbeitsatmosphäre beeindruckt”, erzählt Vasiliy Chitanov, vom bulgarischen zentralem Laboratorium für angewandte Physik, das der bulgarischen Akademie der Wissenschaften angehört. “Ich lernte Michael Krieger von der RISC Software kennen und wir schmiedeten die Idee einer Zusammenarbeit unserer Institutionen. Die RISC Software GmbH hat eine große Expertise im Bereich des Grid Computings, die wir uns für unsere Forschung im Bereich der harten und superharten Beschichtungen zu Nutze machen können.” erklärt Vasiliy. Als 2012 ein entsprechender Fördertopf für bilaterale Zusammenarbeit geöffnet wurde, reichten sie einen Antrag ein und bekamen prompt die Zusage. ” Für RISC Software ist es eine perfekte Gelegenheit, zu zeigen, wie unterschiedliche Analysemethoden von spezifischen Rechenressourcen profitieren können und wie der einzigartige Hintergrund der Mathematik und Informatik auf ein neues Forschungsfeld angewendet werden kann.” Konkret heißt das: es wurden sogenannte virtuelle Maschinen aufgesetzt, mit deren Rechnerleistung die Projektergebnisse des bulgarischen Partners verifiziert werden können.

Unter Grid Computing versteht man eine Form des verteilten Rechnens. Dabei wird die Rechenleistung von virtuellen Maschinen genutzt und die einzelnen Rechenaufgaben verteilt. Das Wort “Grid” ist Englisch und bedeutet auf Deutsch Gitter oder Netz, dies beschreibt die Aufteilung der Rechenleistung auf mehrere Rechenmaschinen.

Die technisch wissenschaftliche Zusammenarbeit zwischen dem bulgarischen zentralen Laboratorium für angewandte Physik und der österreichischen RISC Software GmbH startete vor ungefähr vier Jahren, wurde vom Österreichischen Austauschdienst (oead) gefördert und hatte das Ziel herauszufinden, wie die Experimente des Zentrallaboratoriums für Angewandte Physik am meisten vom Zugang zu effizienten Rechenressourcen profitieren können.

Original Interview in Englisch:

RISC: Please introduce yourself:

Vasiliy: My name is Vasiliy Chitanov. I was born in Plovdiv, which is the second largest town in Bulgaria. I am a researcher at the Central Laboratory of Applied Physics, which is part of the Bulgarian Academy of Sciences. More than 10 years ago, the laboratory has started to focus on the development and transfer of technologies to industry in the area of deposition and characterisation of hard and superhard coatings. My development as researcher at the lab started under the supervision of Assoc. Prof. Dr. Lilyana Kolaklieva (director of the laboratory at the moment) assisting her research work in the development of technologies for metallisation of ohmic contacts. As part of the project’s team in CLAP my next activity, which I am still involved in, is the participation in the characterisation process of coatings developed at CLAP and also of samples from partner organisations and companies. I am working as practical researcher, technologist, and supporting engineer in unbalanced magnetron sputtering deposition techniques. Regarding my education, I am a process control and computer engineer.

Michael: My name is Michael Krieger and I am heading the Advanced Computing Technologies (ACT) department at RISC Software GmbH. In the last eight years, ACT was mainly focusing on optimising software solutions relying on Cloud, HPC, and Grid technologies. Especially in recent years, a growing number of projects have started to focus on big data analysis, data stream processing, and machine learning. Currently we are working on solutions combining the distributed computing background with advanced automated data analysis methods.

RISC: Tell us more about the project you two are currently working on:

Vasiliy: As a participant in the 5th summer school on symbolic calculations in 2010 organised from RISC I was impressed from the infrastructure and the nice working atmosphere in the campus. During my stay, I had the chance to meet Michael Krieger. Following my education and Ph.D. background, I had the idea to establish a new collaboration in the area of GRID computing and software applications in which RISC Software has rich expertise. In 2012 the National Bulgarian Research Fund opened a concourse for bilateral collaboration between Bulgaria and Austria. Together with Michael Krieger we participated with a proposal and we obtained financial support for our project.

Michael: We started our collaboration about 4 years ago. The purpose of a scientific technical cooperation is to start new collaborations on shared research topics. In the framework of this scientific technical cooperation, we have the possibility to investigate how the experiments conducted by the Central Laboratory of Applied Physics can profit most from access to efficient computing resources. For RISC Software it is a perfect opportunity to demonstrate how different analysis methods can profit from specific computing resources and how the unique background of mathematics and computer science can be applied to a new field of research.

RISC: What are the main project goals and what are the main achievements so far?

Vasiliy: There are two main goals of the project. The first one is the development of a small pilot infrastructure in CLAP on which the specialist of CLAP are able to obtain knowledge about GRID computing systems. The goal is that even after the project this infrastructure can be continuously used for research in the area of modelling and simulation of problems form the area of applied physics. The second one is to apply applications and methods developed at RISC and use them as basis for adapting the created pilot GRID infrastructure at CLAP to the needs of applied physics specialists.

Michael: During the first year of the project, we set up a test GRID infrastructure. We used several virtual machines – one as certificate authority, one for job executions and one for user access. The virtual machines were connected with a cloud system hosted at RISC to complete the pilot infrastructure. The different GRID technologies for obtaining certificates, executing GRID jobs and authorized access were applied. Educational video materials were created, which can be used by any new specialist having interest in GRID computing.

Vasiliy: Now the second period of the project has started. During the second year of the project, we will take a closer look at how GRID applications are developed at RISC. A good example is one application for traffic monitoring and analyses. Besides this, we will be investigating possibilities of applying deep learning algorithms to data obtained from the real research experiments in CLAP. The target is to put such applications for data analysis on the pilot GRID infrastructure, which then could be constantly improved and developed for supporting the future research of the CLAP specialists.

I would like to thank Michael Krieger and his co-workers for the possibility to work together on this project. The modelling, simulation and software analysis of deposition PVD processes and data is a very big research topic – it is connected with expensive experiments for validation of the results, characterization of real samples, access to big infrastructures and also collaboration with different specialist from plasma and particle physics, surface chemistry, applied mathematics, vacuum engineering, material science, etc. All this provides the potential for participating in European projects and establish new collaborations in which I believe and hope that RISC and CLAP will continue to collaborate in the future building on the established successful collaboration in this project.

RISC: Thank you for the interview!