Συντάχθηκε 08-08-2012 14:07
από Garantonaki Argyro
Email συντάκτη:
Ενημερώθηκε:
-
Ιδιότητα: -.
ΠΟΛΥΤΕΧΝΕΙΟ ΚΡΗΤΗΣ
Τμήμα Ηλεκτρονικών Μηχανικών & Μηχανικών Υπολογιστών
ΠΑΡΟΥΣΙΑΣΗ ΔΙΠΛΩΜΑΤΙΚΗΣ ΕΡΓΑΣΙΑΣ
Γεωργίας Κουτσανδριά
με θέμα
“Investigation of Resource Allocation and Bandwidth Pricing for Cloud Computing Architectures”
[Τετάρτη 22 Αυγούστου 2012, 11:30 πμ]
[Αίθουσα 145.Π42, Κτίριο Επιστημών, Πολυτεχνειούπολη]
Εξεταστική Επιτροπή
Επ. Καθ. Πολυχρόνης Κουτσακής (επιβλέπων)
Καθ. Μιχάλης Πατεράκης
Καθ. Μίνως Γαροφαλάκης
Abstract
The dramatic increase in cheap broadband wireless and wired network bandwidth along with the drop in storage pricing has lead to the practical fulfillment of computing as utility.
Our work investigates resource allocation strategies for single cloud and multiple distributed clouds under different types of time-varying traffic. A service provider has to assign a request/service from a cloud consumer to the virtual machine that maximizes the profit. Our study focuses on multiple requests/services from a number of consumers and a set of services offered by a particular provider. The revenue of a provider for a particular time frame is defined as the total of values charged to consumers for processing their applications (service requests) during that frame.
We first implement two-profit driven scheduling algorithms from the literature (MaxProfit and MaxUtil) on the well-known and widely used CloudSim simulation tool. We discuss the minor differences in the algorithms’ implementation that were dictated by the CloudSim structure. Then, we propose a new profit-driven algorithm, based on virtual machine CPU utilization, which we compare with MaxProfit and MaxUtil, and we discuss the results of this comparison.
Συνημμένα:
-
DiplomatikiErgasia-Anakoinwsi-Koutsandria.doc
Μεταφορτώσεις: 140,
Μέγεθος: 43 KB application/octet-stream
