Aristotle University Of Thessaloniki (AUTH)
Χρονοδρομολόγηση εργασιών σε περιβάλλοντα υπολογιστικών νεφών
Abstract
dc:descriptionIn today's world, the complex problems that applications have to solve, have exponentially increase the need for computing resources. Cloud computing systems in recent decades have increasingly become the predominant choice by companies and organizations. This is mainly due to the theoretically infinite computing resources that such a system has, direct access from anywhere, fault tolerance and scalability. This rise of Cloud Computing has led providers to offer more and more computing resources to end users as well as sophisticated solutions to meet every need: from data storage and processing to the combined use of multiple computing resources. In addition, the need to avoid increased network traffic and to reduce the time required to upload jobs has led to the invention of new paradigms such as the Fog Computing, the Jungle Computing and a new application development architecture, Microservices.Cloud computing systems were built on top of virtualisation technologies and to this day are inextricably linked to them. There is a plethora of virtualisation technologies that offer the possibility of simultaneous use of computing resources by multiple users, balancing workloads and reducing energy consumption. To achieve the above, Virtual Machines were created. The major benefit of the above implementation is simultaneous use of computing resources without the need for sophisticated knowledge from the end user, since from her view it is "one" big system. The biggest disadvantage of a Virtual Machine is the long start-up time. A solution to this problem was provided by a lighter virtualization technology, Containers. Their key feature is that they make use of the operating system of the real system while they remain isolated, offering better utilization of available resources and reduced boot times.From the above it can be seen that there is a need for better and more flexible scheduling algorithms that can adapt to the large heterogeneity of available computing resources. The aim of this thesis is to analyze and expand the available algorithms, exploring their adaptability to such complex systems in order to balance workload and reduce costs. The systems which are modelled consist of computational resources with a high degree of heterogeneity. Moreover simple jobs and Bag-of-Tasks jobs are used. In addition to this, the examples of Fog Computing, Jungle Computing as well as Hybrid Cloud Computing Systems are considered. For all cases a comparison is made with existing methodologies that are widely used.In summary, the contribution of this thesis lies in the development of task scheduling methodologies in Cloud Computing environments for various types of jobs, either simple or complex. The performance of these methodologies is performed both on real systems and, more importantly, on models of complex heterogeneous systems using simulation.
Degree
thesis:*- Grantor dc:publisher
- Aristotle University Of Thessaloniki (AUTH)
- Year dc:date
- 2021
Author and committee
dc:creator, dc:contributor.*- Authors dc:creator
-
- Tychalas, Dimitrios
- Τυχάλας, Δημήτριος
Subjects
dc:subject × 14- Υπολογιστικό νέφος
- Υψηλή Διαθεσιμότητα
- Εξισορρόπηση φόρτου
- Προσομοίωση
- Cloud computing
- High availability
- Load balancing
- Simulation
- Επιστήμη Ηλεκτρολόγου Μηχανικού, Ηλεκτρονικού Μηχανικού, Μηχανικού Η/Υ
- Επιστήμες Μηχανικού και Τεχνολογία
- Μηχανική και συστήματα επικοινωνιών, Τηλεπικοινωνίες
- Electrical Engineering, Electronic Engineering, Information Engineering
- Engineering and Technology
- Communication engineering and systems, Telecommunications
Rights
- Language dc:language
- gre
Identifiers
dc:identifier.*- Identifier
- 10.12681/eadd/50403