National Technical University of Athens (NTUA)
Τεχνολογίες υπολογιστικού νέφους με έμφαση στη δυναμική αξιολόγηση των παρεχόμενων υπηρεσιών με βάση την ανάλυση της απόδοσης των εφαρμογών και της συγκριτικής αξιολόγησης
Abstract
dc:descriptionCloud services have emerged as an innovative IT provisioning model in the recent years. However, after their usage severe considerations have emerged with regard to their varying performance due to multitenancy and resource sharing issues. These issues make it very difficult to provide any kind of performance estimation during application design or deployment time. In that frame and considering the heterogeneous technology offer and the pricing models currently available in the cloud market a novel mechanism for finding the deployment that fits the application requirements and provides the best Quality of Service (QoS) and cost trade-offs is presented. Based on this a complete definition of cloud performance is presented and various relative subjects are analysed such as the benchmarking of Cloud Providers and the application computational analysis. This thesis focuses on the performance of Cloud services and seeks to introduce and analyze innovative mechanisms for the effective selection of a cloud provider that offers the best environment to host an arbitrary application in terms of performance and cost. On the one hand, the way in which a migrated application component uses the available computing resources may not be known and it might be difficult to choose the best Cloud provider for the component. On the other hand, are Cloud providers interested in knowing the application types that run in their data center in order to avoid interference effects oc concurrently running VMs on performance. So, the current thesis addresses these issues by presenting the mechanisms and methods which identify the computational profile of an arbitrary application component, classify it according to a known limited number of application categories, each represented by a relevant benchmark and provide the best matching with a cloud service solution in terms of performance and cost.Consequently, the experimental results are provided that demonstrate and evaluate the performance and effectiveness of the proposed mechanism for two real application scenarios. Given the experimental results, it appears that the mechanism maps the application profile to a predefined benchmark category and as a result defines the concrete cloud service taking into account the user preferences in terms of performance and cost. The experiments showed promising results and therefore the performance of the mechanism is considered to be well established allowing the adoption of it as a cloud service that seeks to bring QoS knowledge.In addition, the thesis presents a combined methodology which provides an integrated solution for the design and the migration of enterprise applications to Cloud, proving that our innovative mechanism can work complementary with analytical modeling and desing space exploration approaches.Furthermore, the thesis analyzes the performance interference between concurrently running virtual resources and applications in Clouds environments. The same set of application benchmarks as used for Cloud Providers’ benchmarking is chosen that depicts characteristic usage of the hardware resources and the degradation of their performance is measured due to interference effects of concurrently running VM in the same physical host. The outcomes of this experimental process can be used by Cloud Providers as a priori knowledge of theoverhead inserted by the execution of a specific task combination on a physical host.
Degree
thesis:*- Grantor dc:publisher
- National Technical University of Athens (NTUA)
- Year dc:date
- 2017
Author and committee
dc:creator, dc:contributor.*- Authors dc:creator
-
- Evangelinou, Athanasia-Charalampia
- Ευαγγελινού, Αθανασία-Χαραλαμπία
Subjects
dc:subject × 14- Υπολογιστικό νέφος
- υπηρεσίες νέφους
- Βελτιστοποίηση απόδοσης
- Συγκριτική αξιολόγηση
- ανάλυση εφαρμογής
- Cloud computing
- cloud services
- Performance optimization
- Benchmarking
- Application analysis
- Επιστήμες Μηχανικού και Τεχνολογία
- Επιστήμη Ηλεκτρολόγου Μηχανικού, Ηλεκτρονικού Μηχανικού, Μηχανικού Η/Υ
- Engineering and Technology
- Electrical Engineering, Electronic Engineering, Information Engineering
Rights
- Language dc:language
- gre
Identifiers
dc:identifier.*- Identifier
- 10.12681/eadd/40453