logo SBA

ETD

Digital archive of theses discussed at the University of Pisa

 

Thesis etd-07152022-105546


Thesis type
Tesi di dottorato di ricerca
Author
BOLETTIERI, SIMONE
URN
etd-07152022-105546
Thesis title
EDGE-ASSISTED QOS-AWARE RESOURCE AND DATA MANAGEMENT SOLUTIONS FOR IOT APPLICATIONS
Academic discipline
ING-INF/05
Course of study
INGEGNERIA DELL'INFORMAZIONE
Supervisors
tutor Prof. Mingozzi, Enzo
tutor Ing. Bruno, Raffaele
Keywords
  • data management
  • edge
  • edge computing
  • iot
  • mec
  • mobile edge computing
  • optimisation.
  • qos
  • sensor
  • sensor networks
  • service placement
  • ssn
  • wsn
Graduation session start date
25/07/2022
Availability
Withheld
Release date
25/07/2062
Summary
In this thesis, we address the optimal resource and data management problem in heterogeneous and edge-assisted IoT sensing platforms. In these platforms, heterogeneous sensing nodes concurrently feed data to multiple IoT applications and IoT management and operational platforms are deployed at the network edge. In this context, we propose a QoS-aware IoT data brokering solution to perform the optimal: i) collection and multiplex of IoT traffic, ii) edge resource allocation, and iii) data routing between data producers and processing tasks, while maximising QoS requirements of competing and heterogeneous IoT applications. As a key novelty, our broker features caching to facilitate sensor data sharing and improve system scalability. We provide both a mathematical formulation of the data brokering problem and a system architecture. We address the cases in which the broker functionalities are: i) implemented in a single edge device, and ii) distributed on multiple edge nodes. In the latter scenario we jointly tackle the optimisation of the data collection and the allocation of processing tasks to edge nodes. Our design leverages the ETSI MEC standard and entails MEC-compliant components for implementing the distributed brokering platform. Finally, we investigate how a sliced MEC environment can support IoT data management and resource allocation solutions, and we propose a novel slicing architecture that integrates the 3GPP network slicing framework into the MEC infrastructure.
File