Ημερολόγιο Εκδηλώσεων

Περιγραφή:
ΠΟΛΥΤΕΧΝΕΙΟ ΚΡΗΤΗΣ Σχολή Ηλεκτρολόγων Μηχανικών και Μηχανικών Υπολογιστών Πρόγραμμα Προπτυχιακών Σπουδών ΠΑΡΟΥΣΙΑΣΗ ΔΙΠΛΩΜΑΤΙΚΗΣ ΕΡΓΑΣΙΑΣ ΕΜΜΑΝΟΥΗΛ ΣΟΦΙΚΙΤΗ με θέμα Χρήση Τεχνικών Αποθήκευσης Προσωρινής Μνήμης στις Συσκευές των Χρηστών για την Ελάφρυνση Φορτίου Τελικής Ζεύξης σε Δίκτυα Αρχιτεκτονικής C-RAN A Device-to-Device Caching Approach for the Alleviation of the Fronthaul Link Traffic in Centralized Radio Access Networks Εξεταστική Επιτροπή Καθηγητής Μιχαήλ Πατεράκης (επιβλέπων) Καθηγητής Αθανάσιος Λιάβας Αναπληρωτής Καθηγητής Άγγελος Μπλέτσας Abstract The proliferation of modern mobile devices and the ease of access they offer to the Internet, has led to a burst of traffic through the cellular network. This fact leads to the need and creates the motivation for the evolution of Radio Access Networks. Many researchers and organizations suggest the Centralized Radio Access Network architecture as the successor of the current deployment, even though efforts are still ongoing in order to facilitate such architecture, regarding the alleviation of the huge load that would burden the link between the centralized network processors and the antennas deployed throughout the cellular network. This link is referred to as the fronthaul. Considering the ever-growing demand of multimedia services, users’ device level caching appears to be a very promising technique to leverage the traffic offloading of the fronthaul link, while Device-to-Device technology offers the services and functions which facilitate a distribution network between cellular devices. In this thesis, we propose a Device-to-Device caching approach integrated into the Centralized Radio Access Network. According to the proposed approach, some users are storing video files in the local memory of their cellular devices in order to serve video file requests of nearby users. We first describe the main characteristics of the Device-to-Device caching network we consider and we present a system model along with its most important specifications and parameters. Then, we present four case models that we design by gradually relaxing our simplifying system model assumptions and progressively employing improved algorithmic techniques. The performance metric adopted is video hit ratio, since it provides a direct measure of the degree of fronthaul link offloading. Finally, we evaluate and discuss the performance of the above four case models obtained via simulation.