Trial lecture
“Utility function modelling and optimisation for edge computing environments”
Time and place: April 30, 2024 11:15 AM, Kristen Nygaards sal (5370), Ole-Johan Dahls hus
Main research findings
Cloud computing has revolutionized how distributed applications are developed, deployed, and managed. As organizations increasingly rely on Cloud services, optimizing the allocation of resources and the Cloud application configuration to ensure good performance and cost-effectiveness has become a critical challenge. This PhD thesis builds on autonomic computing concepts and aims to enhance the management of Cloud applications in terms of performance, cost, and general user experience. The main topics of this PhD thesis are the modelling of utility functions for Cloud applications and the methods for proactive autonomic Cloud application optimization. Specifically, the thesis introduces novel methods based on templates for modelling utility functions. Furthermore, the thesis addresses the challenge of constrained optimal grouping of Cloud application components by mapping the problem to the Coalition Structure Generation problem and improving the best algorithm from this field to take constraints among the components into consideration. Finally, the thesis proposes how autonomic computing concepts can be used for proactive Cloud application management, making substantial contributions to the field of automatic Cloud application management.
Adjudication committee
- Associate Professor Nelly Bencomo, Department of Computer Science, Durham University, UK
- Dr. Patrizio Dazzi, Department of Computer Science, University of Pisa, Italy
- Professor Martin Steffen, Department of Informatics, University of Oslo, Norway
Supervisors
-
Senior Adviser Geir Henrik Horn, Department of Informatics, UiO, Norway
-
Professor Paulo Ferreira, Department of Informatics, UiO, Norway
-
Professor Tomasz Paweł Michalak, Faculty of Mathematics, Computer Science and Mechanics, University of Warsaw, Poland
Chair of defence:
Professor emeritus Dag Langmyhr
Contact information at Department: Mozhdeh Sheibani Harat