E-Locus - Institutional Repository of the University of Crete

Home Search

Results - Details

Search command : Author="Πλεξουσάκης" And Author="Δημήτρης"

Current Record: 22 of 88

[Add to Basket]

Identifier

000423874

Title

Modelling and dynamic selection of adaptation rules for multi-cloud applications / Eleni Politaki; thesis supervisor, Dimitris Plexousakis.

Alternative Title

Μοντελοποίηση και δυναμική επιλογή κανόνων προσαρμογής για εφαρμογές πολλαπλών υπολογιστικών νεφών

Author

Πολιτάκη, Ελένη Ε.

Thesis advisor

Πλεξουσάκης, Δημήτρης

Reviewer

Μαρκάτος, Ευάγγελος
Μαγκούτης, Κώστας

Abstract

Nowadays, Cloud computing adoption has increased geometrically and many users prefer this type of technology to deploy and manage their applications. Today, there is a high number of Cloud providers, offering a great variety of Cloud services to meet users’ demands. Furthermore, some enterprises prefer to deploy their applications in multiple Clouds in order to benefit from this plethora of offerings. Thus, one important challenge for the Multi-Cloud applications related to the dynamicity and uncertainty that even a single Cloud environment exhibits. As such the increasing complexity makes difficult the delivery of a suitable service level to the customers by the providers. Towards this direction, this thesis introduces two new extensions of the CAMEL modelling language, enabling applications to be adapted across multiple Clouds and different abstraction levels. In addition, an algorithm is proposed for the dynamic selection of the most appropriate adaptation rule for each problematic situation, based on its priority, according to the adaptation history of the application. In the first part of this thesis, we focus on the proposed extensions of CAMEL, based on an existing cross-level and Multi-Cloud application adaptation architecture. Adaptation actions, rules and strategies are central adaptation-related notions that played a fundamental role in the extensions performed in the CAMEL meta-model. Adaptation rules match an event or event pattern, representing an occurrence of a critical situation, with adaptation workflows, which specify the concrete adaptation actions to be performed for addressing this situation, while adaptation strategies are necessary both for organizing the set of adaptation rules in the context of the same event or event pattern that triggers them, and for representing the application’s adaptive behavior. In the second part of this thesis, we elaborate on the dynamic selection algorithm of the most appropriate adaptation rule, within an adaptation strategy, which is based on its priority value for addressing a certain problematic situation represented by an event or event pattern. This priority value is calculated on the basis of a specific mathematical formula. In the adaptation history of each application recorded particular sensor measurements which are exploited for the computation of the quality attributes that participate in the priority formula to be calculated and come from previous executions of the adaptation rules selected. Thus, the main contributions of this thesis aim to the better management of the applications that are executed in Multi-Cloud environments by the use of cross-layer adaptation workflows, and the dynamic selection of the most appropriate adaptation rule.

Language

English

Subject

Adaptation rules

Multi-cloud

Κανόνες προσαρμογής

Πολλαπλά νέφη

Issue date

2019-07-26

Collection

School/Department--School of Sciences and Engineering--Department of Computer Science--Post-graduate theses

Type of Work--Post-graduate theses