Your browser does not support JavaScript!

Home    Collections    Type of Work    Post-graduate theses  

Post-graduate theses

Current Record: 4806 of 6549

Back to Results Previous page
Next page
Add to Basket
[Add to Basket]
Identifier 000403635
Title A distributed cross-layer monitoring system based on QoS metrics models
Alternative Title Ένα κατανεμημένο πολυεπίπεδο σύστημα παρακολούθησης βασισμένο σε QoS μετρικά μοντέλα
Author Μεταλλίδης, Δαμιανός Γ.
Thesis advisor Πλεξουσάκης, Δημήτρης
Reviewer Μαρκάτος, Ευάγγελος
Μαγκούτης, Κώστας
Abstract In order to implement cross-organizational workflows and to realize collaborations between small and medium enterprises (SMEs) the use of Web service technology and Service-Oriented Architecture have become a necessity. Whilst, SMEs are continuously moving towards service-oriented infrastructures where applications are being modeled, the need of hosting them has raised an important issue for the quality of the underlying cloud infrastructures. Virtualization, offered by cloud infrastructures, delegates the use of any kind of resources, such as computing environments or storage systems, to the data center's internal networks. All of the above issues raised the need for monitoring of the quality of the acquired resources and of the services offered to final users as also the workflow-based procedures used by SMEs in order to use services. Although, most of the monitoring frameworks and systems rely only on a layer-specific quality model, covering, e.g., the service layer, without considering the cross-layer dependencies it might have with the Quality Models (QMs) of Workflow and Infrastructure. To alleviate this problem in this thesis, we have defined three metric QMs that cover quality terms based on Workflow, Service and Infrastructure layers. These quality terms could refer to (a) quality dimensions describing the quality aspect which can be used to provide an aspect-specific partition of quality terms (b) quality attributes indicating properties of an object (e.g., object of response time) (c) raw quality metrics representing raw information that could be taken from monitoring sensors and (d) composite metrics which can be used in the calculation of different composite metrics by applying a specific metric formula which is defined in order to calculate the value of the according composite metric. The novelty of our approach is the definition of a fourth one QM depicting two types of cross-layer dependencies, between the quality metrics of the aforementioned three QMs, which could be (1) equality relations and (2) inter-dependency relations. Furthermore, in order to compute the values of the quality metrics for each of these three QMs we have defined aggregators, computation formulas, and algorithms which are implemented and placed within a distributed cross-layer monitoring system. Implementation of the cross-layer monitoring system is based upon the two open-source monitoring tools of Prometheus and Nagios, for Service and Infrastructure layers accordingly. As a workflow engine we have used the Activiti open source project, responsible for the execution and monitoring of cross-organizational workflows. We evaluated the proposed cross-layer monitoring system by deploying its components in the private cloud of VMWare and in the public cloud provider of University of Crete Data Center (UCDC). Our evaluation procedure is based on (a) the exposition of a real life workflow example indicating the virtue of the unified dependency quality model proposed and explicates the way computation algorithms work (b) the performance aspect of the proposed cross-layer monitoring system such as the query execution latency upon the management databases being used, in both private and public cloud providers and (c) the value of accuracy offered by the distributed monitoring system.
Language English
Subject Dependency relations
Workflows
Παρακολούθηση
Ροές εργασίας
Σχέσεις εξάρτησης
Issue date 2016-11-18
Collection   School/Department--School of Sciences and Engineering--Department of Computer Science--Post-graduate theses
  Type of Work--Post-graduate theses
Permanent Link https://elocus.lib.uoc.gr//dlib/e/5/7/metadata-dlib-1478336033-255603-16543.tkl Bookmark and Share
Views 599

Digital Documents
No preview available

Download document
View document
Views : 29