Tesi etd-09192023-161747
Link copiato negli appunti
Tipo di tesi
Dottorato
Autore
SCANO, DAVIDE
URN
etd-09192023-161747
Titolo
Programmable Data Plane for Disaggregated Optical Networks
Settore scientifico disciplinare
ING-INF/03
Corso di studi
Istituto di Tecnologie della Comunicazione, dell'Informazione e della Percezione - PHD IN EMERGING DIGITAL TECHNOLOGIES
Commissione
relatore Prof. CASTOLDI, PIERO
Parole chiave
- P4
- SDN
- Networking
- Telemetry
Data inizio appello
14/12/2023;
Disponibilità
completa
Riassunto analitico
Software-defined networking (SDN) is a networking paradigm that offers fine-grained control, flexibility, and openness compared to traditional networks. SDN allows the separation of the control plane from the data plane, providing a global view of the network through a logically centralized control, and enabling tailored networking solutions through a reconfigurable data plane.
In particular, the advent of Programming Protocol-Independent Packet Processors (P4) spearheads the era of SDN-driven data plane reconfigurability via the complete programmability of the packet network.
However, the full programmability of the data plane has mainly been implemented and embraced within data centers, leaving its potential application in metropolitan area networks largely unexplored.
This dissertation delves into the exploitation of programmability to enhance control in SDN-driven disaggregated metro networks operated by Telecom Operators.
Indeed, the advent of 5G and beyond (B5G) and edge computing technologies poses a challenge to disaggregated metro networks as they strive to meet demanding performance requirements. Furthermore, with the increasing convergence of packet and optical technologies, closer coordination between metro and transport networks is necessary.
The novel programmable data plane solutions proposed in this thesis focus on three main scenarios for the evolution of disaggregated metro networks: (i) 5G and beyond, (ii) edge computing, and (iii) coordination between metro and optical transport networks.
With B5G architecture, there's a need to support perceived zero latency applications, which translates into technical requirements for accurate end-to-end (e2e) network control from the cloud to B5G user equipment (UE).
The aim is to exploit network programmability to provide novel decentralized capabilities, including effective pre-planned self-management of low-latency services.
In the context of edge computing, programmability is leveraged to improve service chaining performance and provide effective joint allocation of computation and network resources.
In a first scenario, programmability is employed to enforce proactive in-network functions, such as priority changes, ensuring bounded service function chaining (SFC) and delivering segment latency. In a second scenario, programmability enables fine-grained control and monitoring of tunneled traffic in the data plane. This information is then utilized in the control plane to optimize the allocation of computing resources.
To enhance the closer coordination between metro and transport networks, programmability is synergistically combined with IP over Wavelength Division Multiplexing (IPoWDM) and coherent pluggable modules. This approach facilitates the augmentation of network telemetry with optical parameters, enabling more comprehensive monitoring and in-depth analysis of network performance. Moreover, a novel control architecture is proposed to effectively address the coordination challenges between metro and transport networks, ensuring seamless harmonization and efficient operation.
In particular, the advent of Programming Protocol-Independent Packet Processors (P4) spearheads the era of SDN-driven data plane reconfigurability via the complete programmability of the packet network.
However, the full programmability of the data plane has mainly been implemented and embraced within data centers, leaving its potential application in metropolitan area networks largely unexplored.
This dissertation delves into the exploitation of programmability to enhance control in SDN-driven disaggregated metro networks operated by Telecom Operators.
Indeed, the advent of 5G and beyond (B5G) and edge computing technologies poses a challenge to disaggregated metro networks as they strive to meet demanding performance requirements. Furthermore, with the increasing convergence of packet and optical technologies, closer coordination between metro and transport networks is necessary.
The novel programmable data plane solutions proposed in this thesis focus on three main scenarios for the evolution of disaggregated metro networks: (i) 5G and beyond, (ii) edge computing, and (iii) coordination between metro and optical transport networks.
With B5G architecture, there's a need to support perceived zero latency applications, which translates into technical requirements for accurate end-to-end (e2e) network control from the cloud to B5G user equipment (UE).
The aim is to exploit network programmability to provide novel decentralized capabilities, including effective pre-planned self-management of low-latency services.
In the context of edge computing, programmability is leveraged to improve service chaining performance and provide effective joint allocation of computation and network resources.
In a first scenario, programmability is employed to enforce proactive in-network functions, such as priority changes, ensuring bounded service function chaining (SFC) and delivering segment latency. In a second scenario, programmability enables fine-grained control and monitoring of tunneled traffic in the data plane. This information is then utilized in the control plane to optimize the allocation of computing resources.
To enhance the closer coordination between metro and transport networks, programmability is synergistically combined with IP over Wavelength Division Multiplexing (IPoWDM) and coherent pluggable modules. This approach facilitates the augmentation of network telemetry with optical parameters, enabling more comprehensive monitoring and in-depth analysis of network performance. Moreover, a novel control architecture is proposed to effectively address the coordination challenges between metro and transport networks, ensuring seamless harmonization and efficient operation.
File
| Nome file | Dimensione |
|---|---|
| PhD_Thes...final.pdf | 28.77 Mb |
Contatta l'autore |
|