DTA

Digital Theses Archive

 

Tesi etd-08092019-153356

Type of thesis
Dottorato
Author
CIVERCHIA, FEDERICO
URN
etd-08092019-153356
Title
How will edge computing shape the 5G deployment? The hardware acceleration use case
Structure
ING-INF/05
Course
INGEGNERIA - Ph.D. Programme in Emerging Digital Technologies (EDT)
Committee
Membro Prof. CASTOLDI, PIERO
Presidente Prof. AZCORRA, ARTURO
Membro Prof. PELCAT, MAXIME
Membro Prof. VALCARENGHI, LUCA
Membro Prof. FRANKLIN, ANTONY
Membro Dott. CERRONI, WALTER
Membro Prof. MONTI, PAOLO
Keywords
  • 5G
  • BMV2
  • Cyber Security
  • Edge Node
  • Flowlet
  • FPGA
  • Hardware Acceleration
  • Multi-Layer
  • NetFPGA
  • NFV
  • OpenCL
  • Optical Bypass
  • P4
  • Pipelined Service Chain
  • Reconfigurable Computing
  • SDN
  • SYN Flood
  • Token Bucket
  • Traffic Engineering
Exam session start date
;
Availability
completa
Abstract
The recent improvements in the information technology will lead to the new era of the communication (5G) where everything will be connected, where smart and connected objects will be a constant presence in our daily life. Thus, stricter requirements in terms of communication bandwidth and latency have to be satisfied to meet the demands of a huge number of connected devices. Instead of the centralized approach, moving the data processing to the edge can improve the performance since it reduces the infrastructure-user round trip time and it saves Cloud bandwidth. However, this decentralization comes at a price though. Moving data computation and communication processing to the network edges improves system scalability and reliability but it requires more local hardware and only a subset of data is analyzed. This means that an edge system does not have global visibility of the information.<br>This thesis aims at presenting a novel approach to accelerate the 5G infrastructure at the edge. The idea is to exploit hardware acceleration to improve the processing of the protocol stack functionalities and network functions close to the final user. In this way, the bandwidth for the communication between 5G radio infrastructure and the central Cloud can be saved. Moreover, real time application can benefit from the improved computation capabilities by means of hardware offloading.<br>Considering the latest developments in the embedded systems in terms of computational power and lower hardware cost, we envision that edge computing can be exploited to improve 5G infrastructure. Thus, the edge computing is ready to be deployed in 5G architecture, improving the user experience.
Files