Cloud Computing

The Cloud Computing paradigm holds good promise for the performance hungry scientific community. Clouds promise to be a cheap alternative to supercomputers and specialized clusters, a much more reliable platform than grids, and a much more scalable platform than the largest of commodity clusters or resource pools. Clouds also promise to “scale by credit card”, that is, scale up immediately and temporarily with the only limits imposed by financial reasons, as opposed to the physical limits of adding nodes to clusters or even supercomputers or to the financial burden of over-provisioning resources.

cloud-architecture

Many Cloud computing companies distinguish themselves in the type of services they offer. At the highest-level, we observe three main directions:

  • Infrastructure as a Service (IaaS)
  • Software as a Service (SaaS)
  • Platform as a Service (PaaS)

 

 

The DPS group is focusing on the following criteria in their Cloud computing related research and most of these approaches are integrated into the ASKALON system:

Hardware and Software Virtualisation

Does virtualisation show side effects on scientific computing? How can this technology be used to optimize scheduling, minimize load imbalance and what are other ways to benefit from this hardware abstraction layer?

Performance modeling and analysis

Is the performance of Clouds sufficient for scientific high performance computing? The information provided by commercial Cloud providers is not specific enough for precise scheduling and planning decisions. The offered performance needs to be modeled and analyzed to match the requirements of planned executions.

Quality of service / Service-level agreements

How can Clouds offer Quality of Service (QoS) guarantees? Cloud Computing relies on highly distributed systems, and starting from the network connection there is the question: Can the provider guarantee that the resources he sells will be accessible by the user? Therefore SLA negotiation, automation and orchestration is needed in Cloud environments.

Multi-criteria Scheduling

Running scientific workflow applications on Cloud infrastructures requires careful consideration of multiple criteria during the scheduling process. As an example, it is usually not sufficient to schedule the application just for minimum execution time, because this can drive the costs too high, especially on commercial Cloud systems.

RainCloud Project

The DPS group is part of the RainCloud project done together with Institute of Meteorology and Geophysics at the University of Innsbruck: http://imgi.uibk.ac.at/groups/raincloud.

Power Management

Measureing power consumption on hardware and optimizing the system utilization to minimize the power consumption as well is also covered by the research done by the Cloud division of the DPS group.

Staff

  • Priv.-Doz. Dr. Radu Prodan
  • Prof. Dr. Thomas Fahringer
  • Dr. Simon Ostermann
  • Dr. Hamid Mohammadi Fard
  • Matthias Janetschek

Selected Publications

  • Radu Prodan, Michael Sperk, Simon Ostermann. Experimental Approach to Evaluate HPC on Google App Engine, IEEE Software, url
  • Simon Ostermann, Kassian Plankensteiner and Radu Prodan. Using a New Event-based Simulation Framework for Investigating Different Resource Provisioning Methods in Clouds, Scientific Programming Journal, Volume 19, Number 2-3 / 2011, pages 161-178, DOI: 10.3233/SPR-2011-0321 url
  • Alexandru Iosup, Simon Ostermann, Nezih Yigitbasi, Radu Prodan, Thomas Fahringer and Dick Epema, Performance Analysis of Cloud Computing Services for Many-Tasks Scientific Computing, IEEE Transactions on Parallel and Distributed, Systems Special Issue on Many-Task Computing, JUNE 2011 (Vol. 22, No. 6) pp. 931-945, url
  • Vlad Nae, Radu Prodan, and Alexandru Iosup, Massively Multiplayer Online Game Hosting on Cloud Resources, Cloud Computing: Principles and Paradigms, 2009, ch. 23, ed. R. Buyya, j. Broberg, A. Goscinski, Wiley STM
  • Simon Ostermann, Radu Prodan, and Thomas Fahringer, Resource Management for Hybrid Grid and Cloud Computing, Cloud Computing: Principles, Systems and Applications, Springer, 2010. chapter 11, page 179-194. DOI: 10.1007/978-1-84996-241-4, ISBN 1849962405url
  • Gabriela Andreea Morar, Felix Schueller, Simon Ostermann, Radu Prodan and Georg Mayr. Meteorological Simulations in the Cloud with the ASKALON Environment. In CoreGRID/ERCIM Workshop on Grids, Clouds and P2P Computing (CGWS 2012), Rhodes Island, August 27, 2012 url
  • Simon Ostermann, Radu Prodan, Impact of Variable Priced Cloud Resources on Scientific Workflow Scheduling . Accepted for publication in Euro-Par 2012, International European Conference on Parallel and Distributed Computing, Rhodes Island, Greece, August 27th – August 31st, 2012, page 350-362, url
  • Hamid Mohammadi Fard, Radu Prodan, Thomas Fahringer, Juan Jose Durillo Barrionuevo, A Multi-Objective Approach for Workflow Scheduling in Heterogeneous Computing Environments to appear in CCGrid 2012 — The 12th IEEE/ACM International Symposium on Cluster, Cloud and Grid Computing, May 13-16, 2012, Ottawa, Canada. url
  • Simon Ostermann, Radu Prodan and Thomas Fahringer, Dynamic Cloud Provisioning for Scientific Grid Workflows, in The 11th ACM/IEEE International Conference on Grid Computing (Grid 2010), Brussels, Belgium, October 25 – 29, 2010, ISBN 978-1-4244-9348-7, Pages 97-104.
  • Simon Ostermann, Kassian Plankensteiner, Daniel Bodner, Georg Kraler and Radu Prodan, Integration of an event-based simulation framework into a scientific workflow execution environment for Grids and Clouds, 4th European Conference ServiceWave 2011, Polzan, Poland, 26-28 October, 2011. url
  • Simon Ostermann, Kassian Plankensteiner, Radu Prodan and Thomas Fahringer, GroudSim: An Event-based Simulation Framework for Computational Grids and Clouds, To be published in: CoreGRID/ERCIM Workshop on Grids, Clouds and P2P Computing in conjunction with EuroPAR 2010, Ischia, Naples, Italy, August 31, 2010
  • Vlad Nae and Radu Prodan and Thomas Fahringer, Cost-Efficient Hosting and Load Balancing of Massively Multiplayer Online Games, Proceedings of the 11th IEEE/ACM International Conference on Grid Computing (Grid), 2010, p. 9-17, IEEE Computer Society
  • Alexandru Iosup and Vlad Nae and Radu Prodan, The Impact of Virtualization on the Performance and Operational Costs of Massively Multiplayer Online Games, International Journal of Advanced Media and Communication (IJAMC), 2011, vol. 4, p. 364-386
  • Vlad Nae, Alexandru Iosup, Radu Prodan, Dynamic Resource Provisioning in Massively Multiplayer Online Games, IEEE Transactions on Parallel and Distributed Systems (TPDS), 2010, vol. PP, issue 99
  • Radu Prodan, Vlad Nae, edutain@grid: A Resource Management Platform for Massively Multiplayer Online Games, Computer Games: Learning Objectives, Cognitive Performance and Effects on Development, 2010, p. 1–51, ed. Agustin Soria and Julián Maldonad, Nova Science Publishers
  • Vlad Nae, Radu Prodan, Thomas Fahringer, Monitoring and Fault Tolerance for Real-Time Online Interactive Applications, Euro-Par 2009 — Parallel Processing Workshops, Lecture Notes in Computer Science, 2010, vol. 6043, p. 255-265, Springer Berlin – Heidelberg
  • Vlad Nae, Alexandru Iosup, Radu Prodan, Thomas Fahringer, The Impact of Virtualization on the Performance of Massively Multiplayer Online Games, 8th Annual Workshop on Network and Systems Support for Games (NetGames), 2009, p. 1 – 6
  • Radu Prodan, Vlad Nae, Thomas Fahringer and Herbert Jordan, Dynamic Real-Time Resource Provisioning for Massively Multiplayer Online Games, Parallel Computing Technologies, Lecture Notes in Computer Science, 2009, vol. 5698, p. 98-111, Springer Berlin — Heidelberg
  • Vlad Nae, Radu Prodan, and Alexandru Iosup, Massively Multiplayer Online Game Hosting on Cloud Resources, Cloud Computing: Principles and Paradigms, 2009, p. 491, ed. R. Buyya, j. Broberg, A. Goscinski, Wiley STM
  • Radu Prodan and Vlad Nae, Prediction-based Real-Time Resource Provisioning for Massively Multiplayer Online Games, Future Generation Computer Systems (FGCS), 2009, vol. 25, issue 7, p. 785-793
  • Vlad Nae, Jordan Herbert, Radu Prodan, and Thomas Fahringer, An Information System for Real-Time Online Interactive Applications, Euro-Par 2008 Workshops – Parallel Processing, Lecture Notes in Computer Science, 2009, vol. 5415, p. 361-370, Springer Berlin, Heidelberg
  • Vlad Nae, Alexandru Iosup, Stefan Podlipnig, Radu Prodan, Dick Epema, Thomas Fahringer, Efficient Management of Data Center Resources for Massively Multiplayer Online Games, SC ’08: Proceedings of the 2008 ACM/IEEE conference on Supercomputing, 2008, p. 1–12, IEEE Press
  • Simon Ostermann, Radu Prodan, and Thomas Fahringer, Extended Grids with Cloud Resource Management for Scientific Computing, Cloud Computing: Principles, Systems and Applications, Springer, 2010. Accepted for publication.
  • Simon Ostermann, Alexandru Iosup, Nezih Yigitbasi, Radu Prodan, Thomas Fahringer and Dick Epema, A Performance Analysis of EC2 Cloud Computing Services for Scientific Computing. In Proceedings of Cloudcomp 2009 (CloudComp), October 19-21, 2009, Munich, Germany. ISBN 978-963-9799-77-6, Page x-(x 17).
  • Radu Prodan and Simon Ostermann, A Survey and Taxonomy of Infrastructure as a Service and Web Hosting Cloud Providers. In Proceedings of The 10th IEEE/ACM International Conference on Grid Computing (Grid 2009), October 13-15, 2009, Banff, Canada. ISBN 978-1-4244-5149-4, Page 17-25.
  • Kassian Plankensteiner, Radu Prodan, Thomas Fahringer, A New Fault Tolerance Heuristic for Scientific Workflows in Highly Distributed Environments based on Resubmission Impact In Proceedings of the 5th IEEE International Conference on e-Science, December 09-11, 2009, Oxford, UK. Accepted for publication.
  • Simon Ostermann, Radu Prodan and Thomas Fahringer, Extended Grids with Cloud Resource Management for Scientific Computing. In Proceedings of The 10th IEEE/ACM International Conference on Grid Computing (Grid 2009), October 13-15, 2009, Banff, Canada. ISBN 978-1-4244-5149-4, Page 42-59. bib
  • Vlad Nae, Radu Prodan, Thomas Fahringer, Neural Network-based Load Prediction for Highly Dynamic Distributed Online Games. in Proceedings of EuroPar 2008, ISBN: 978-3-540-85450-0, Page 202-211, Canary Island, Spain, August 26-29, 2008.
  • Nezih Yigitbasi, Alexandru Iosup, Simon Ostermann, and Dick Epema. C-Meter: A Framework for Performance Analysis of Computing Clouds. In the International Workshop on Cloud Computing (Cloud 2009), May 18-21, 2009, in conjunction with CCGrid’09, ISBN 978-1-4244-3935-5, Page 472-477 bib
  • Simon Ostermann, Ressource Management for Scientific Application in Hybrid Cloud Computing Environments, University of Innsbruck, Austria, Institute of Computer Science, PhD thesis, 2012.