Optical networks represent the essential backbone for various communication systems, such as long-haul, metro, and data center networks. As future optical networks will have to handle a dramatic increase in the volume of highly sensitive data, it is crucial for future optical networks to energy-efficiently accommodate increasing traffic demands as well as to be resilient against network failures and secure against attackers. Specifically, as current communication systems already contribute 2% to 3% of global energy consumption, and with this figure expected to rise due to increasing traffic demands in future optical networks, it is crucial to develop sustainable solutions to reduce energy consumption. Moreover, ensuring network resiliency during failures, such as those caused by natural disasters, is vital for maintaining uninterrupted services. At the same time, quantum computers present significant security challenges to current cryptosystems, making it essential to develop effective countermeasures against potential quantum attacks. This thesis aims to design resource allocation algorithms to improve the energy efficiency, resiliency, and security of future optical networks, and the main contributions can be summarized into three parts as follows. (1) Proposal to Improve Energy Efficiency based on Novel Optical-Transmission Technologies: Recent advances in optical-transmission technologies bring new possibilities for enhancing energy efficiency for future optical networks. I have designed resource allocation algorithms for two emerging optical-transmission approaches, namely pluggable optics and power profile monitoring (PPM). Regarding pluggable optics, I have comprehensively investigated the energy efficiency of different ''IP over Wavelength Division Multiplexing'' (IPoWDM) network architectures with ZR/ZR+, and evaluated the advantages of ZR/ZR+ compared to long-haul muxponders. Specifically, I have proposed a pragmatic power consumption model for optical nodes and designed an auxiliary-graph-based approach to quantify the power consumption of different IPoWDM network architectures. Regarding PPM, I have proposed and investigated the novel optimized monitoring placement (OMP) problem for PPM to minimize the monitoring power consumption. Then, I have quantitatively compared the power consumption of PPM vs. the traditional monitoring technique, Optical Time-Domain Reflectometer (OTDR), with both an Integer Linear Programming (ILP) model for small-scale scenarios and an efficient heuristic algorithm for large-scale scenarios. These works provide guidelines for the deployment of ZR/ZR+ and PPM to improve the energy efficiency of future optical networks. (2) Proposal to Enhance Network Resiliency with Proactive and Reactive Approaches: A key challenge in network virtualization for future optical networks is to improve network resiliency against possible failures and enable swift recovery after failures. I have investigated proactive solutions to improve network resiliency against double-link failures and reactive solutions to provide swift network recovery under massive failures. Specifically, to provide network resiliency against any possible double-link failures in the network, I have investigated capacity sharing among existing virtual networks to reduce the additional resources to enhance network resiliency. More specifically, I have explored two strategies to enhance resiliency through capacity sharing, one employing a dedicated slice to improve resiliency and the other without such a slice. To solve this problem, I have formulated an ILP model and designed a scalable local-search-based heuristic algorithm. Regarding swift recovery, I have introduced a novel problem, called progressive slice recovery (PSR) with guaranteed reachability among network nodes (network connectivity) and reachability of data centers (content connectivity). Specifically, I have investigated to determine a recovery sequence for slices to swiftly recover the slices and, in the meantime, consider accelerating the provision of network services with connectivity constraints. To solve the PSR problem, I have developed an ILP model and a two-phase progressive slice recovery algorithm based on column generation and deterministic rounding. The numerical results confirmed that network resiliency can be guaranteed with less additional resources through proactive approaches and shorter recovery time using reactive approaches. (3) Proposal to Improve Network Security with Quantum Technologies: Quantum computers pose emerging security threats to future optical networks, making it critical to explore novel algorithms and architectures with quantum technologies. I have investigated novel resource allocation algorithms and architectures for QKD and quantum networks, enhancing their ability to securely distribute keys among nodes with guaranteed information-theoretic security. First, I have introduced a novel problem of Routing, Channel, Key-rate and Time-slot Assignment (RCKTA) to allocate quantum resources for QKD networks. This problem considers incorporating the advantages of trusted relays to increase the key rate and optical bypassing to reduce the unnecessary QKD modules (i.e., transceivers for QKD networks). To solve the RCKTA problem, I have devised a Mixed Integer Linear Programming (MILP) model and a near-optimal heuristic algorithm, achieving a substantial increase in acceptance ratio. Next, I have designed a novel QKD network architecture integrating Information-Centric Networking (ICN) to address the challenges of limited secret key rates and diverse service requirements. This architecture can accelerate key distribution with in-network key caching of ICN and meet diverse service requirements by analyzing the requirements based on semantic information. Lastly, I have investigated link configuration for fidelity-constrained routing and purification (LC-FCRP) in quantum networks, developing a MILP model and a shortest-path-based fidelity determination algorithm to optimize the fidelity of distributed entanglements. These approaches significantly increase the performance of QKD networks and quantum networks to guarantee the security of future optical networks.
Le reti ottiche rappresentano la spina dorsale essenziale per vari sistemi di comunicazione, come le reti a lunga distanza, metropolitane e i centri dati. Poiché le future reti ottiche dovranno gestire un aumento drammatico del volume di dati altamente sensibili, è cruciale che queste siano in grado di accogliere in modo efficiente dal punto di vista energetico le crescenti richieste di traffico, oltre a essere resilienti contro i guasti di rete e sicure contro gli attacchi. In particolare, poiché i sistemi di comunicazione attuali già contribuiscono dal 2% al 3% del consumo energetico globale, e dato che si prevede un aumento di questa cifra a causa dell'aumento del traffico nelle future reti ottiche, è fondamentale sviluppare soluzioni sostenibili per ridurre il consumo energetico. Inoltre, garantire la resilienza della rete durante i guasti, come quelli causati da disastri naturali, è vitale per mantenere i servizi ininterrotti. Allo stesso tempo, i computer quantistici presentano significative sfide alla sicurezza per i crittosistemi attuali, rendendo essenziale sviluppare contromisure efficaci contro potenziali attacchi quantistici. Questa tesi mira a progettare algoritmi di allocazione delle risorse per migliorare l'efficienza energetica, la resilienza e la sicurezza delle future reti ottiche, e i principali contributi possono essere riassunti in tre parti come segue. (1) Proposta per Migliorare l'Efficienza Energetica basata su Nuove Tecnologie di Trasmissione Ottica: I recenti progressi nelle tecnologie di trasmissione ottica offrono nuove possibilità per migliorare l'efficienza energetica delle future reti ottiche. Ho progettato algoritmi di allocazione delle risorse per due approcci emergenti alla trasmissione ottica, ossia l'ottica pluggabile e il monitoraggio del profilo di potenza (PPM). Per quanto riguarda l'ottica pluggabile, ho investigato approfonditamente l'efficienza energetica di diverse architetture di rete ''IP su Wavelength Division Multiplexing'' (IPoWDM) con ZR/ZR+, e valutato i vantaggi di ZR/ZR+ rispetto ai muxponder a lunga percorrenza. In particolare, ho proposto un modello pragmatico di consumo energetico per i nodi ottici e progettato un approccio basato su grafi ausiliari per quantificare il consumo energetico di diverse architetture di rete IPoWDM. Riguardo al PPM, ho proposto e investigato il nuovo problema ottimizzato di posizionamento del monitoraggio (OMP) per il Power Profile Monitoring (PPM) per minimizzare il consumo energetico del monitoraggio. Poi, ho confrontato quantitativamente il consumo energetico del PPM rispetto alla tecnica di monitoraggio tradizionale, l'Optical Time-Domain Reflectometer (OTDR), con sia un modello di Programmazione Lineare Intera (ILP) per scenari di piccola scala sia un algoritmo euristico efficiente per scenari di grande scala. Questi lavori forniscono linee guida per il dispiegamento di ZR/ZR+ e PPM per migliorare l'efficienza energetica delle future reti ottiche. (2) Proposta per Migliorare la Resilienza della Rete con Approcci Proattivi e Reattivi: Una sfida chiave nella virtualizzazione della rete per le future reti ottiche è migliorare la resilienza della rete contro possibili guasti e abilitare un rapido recupero dopo i guasti. Ho investigato soluzioni proattive per migliorare la resilienza della rete contro i guasti doppi collegamenti e soluzioni reattive per fornire un recupero della rete rapido sotto guasti massivi. Specificamente, per fornire resilienza della rete contro qualsiasi possibile guasto doppio collegamento nella rete, ho investigato la condivisione di capacità tra le reti virtuali esistenti per ridurre le risorse aggiuntive per potenziare la resilienza della rete. Più specificamente, ho esplorato due strategie per migliorare la resilienza tramite la condivisione di capacità, una impiegando una fetta dedicata e l'altra senza tale fetta. Per risolvere questo problema, ho formulato un modello ILP e progettato un algoritmo euristico basato su ricerca locale scalabile. Riguardo al recupero rapido, ho introdotto un nuovo problema, chiamato recupero progressivo della fetta (PSR) con garanzia di raggiungibilità tra i nodi della rete (connettività della rete) e raggiungibilità dei centri dati (connettività dei contenuti). Specificamente, ho investigato per determinare una sequenza di recupero per le fette per recuperarle rapidamente e, nel frattempo, considerare l'accelerazione della fornitura di servizi di rete con vincoli di connettività. Per risolvere il problema PSR, ho sviluppato un modello ILP e un algoritmo di recupero della fetta progressivo in due fasi basato sulla generazione di colonne e arrotondamento deterministico. I risultati numerici hanno confermato che la resilienza della rete può essere garantita con meno risorse aggiuntive per approcci proattivi e tempi di recupero più corti per approcci reattivi. (3) Proposta per Migliorare la Sicurezza della Rete con Tecnologie Quantistiche: I computer quantistici pongono minacce alla sicurezza emergenti per le future reti ottiche, rendendo critico esplorare nuovi algoritmi e architetture con tecnologie quantistiche. Ho investigato nuovi algoritmi di allocazione delle risorse e architetture per QKD e reti quantistiche, migliorando la loro capacità di distribuire chiavi in modo sicuro tra i nodi con sicurezza garantita a livello teorico delle informazioni. Per primo, ho introdotto un nuovo problema di Routing, Channel, Key-rate e Time-slot Assignment (RCKTA) per allocare risorse quantistiche per le reti QKD. Questo problema considera l'incorporazione dei vantaggi dei relay fidati per aumentare la frequenza delle chiavi e il bypass ottico per ridurre i moduli QKD non necessari (ovvero i trasmettitori per le reti QKD). Per risolvere il problema RCKTA, ho ideato un modello di Programmazione Lineare Intera Mista (MILP) e un algoritmo euristico quasi ottimale, ottenendo un sostanziale aumento nel rapporto di accettazione. Successivamente, ho progettato una nuova architettura di rete QKD integrando il Networking Centrato sull'Informazione (ICN) per affrontare le sfide delle limitate frequenze di chiave segreta e dei diversi requisiti di servizio. Questa architettura può accelerare la distribuzione delle chiavi con la memorizzazione delle chiavi in rete di ICN e soddisfare i diversi requisiti di servizio analizzando i requisiti basati su informazioni semantiche. Infine, ho investigato la configurazione dei collegamenti per il routing e la purificazione vincolati dalla fedeltà (LC-FCRP) nelle reti quantistiche, sviluppando un modello MILP e un algoritmo di determinazione della fedeltà basato sul cammino più breve per ottimizzare la fedeltà degli entanglement distribuiti. Questi approcci aumentano significativamente le prestazioni delle reti QKD e delle reti quantistiche per garantire la sicurezza delle future reti ottiche.
Green, resilient, and secure next-generation optical networks
QIAOLUN, ZHANG
2025
Abstract
Optical networks represent the essential backbone for various communication systems, such as long-haul, metro, and data center networks. As future optical networks will have to handle a dramatic increase in the volume of highly sensitive data, it is crucial for future optical networks to energy-efficiently accommodate increasing traffic demands as well as to be resilient against network failures and secure against attackers. Specifically, as current communication systems already contribute 2% to 3% of global energy consumption, and with this figure expected to rise due to increasing traffic demands in future optical networks, it is crucial to develop sustainable solutions to reduce energy consumption. Moreover, ensuring network resiliency during failures, such as those caused by natural disasters, is vital for maintaining uninterrupted services. At the same time, quantum computers present significant security challenges to current cryptosystems, making it essential to develop effective countermeasures against potential quantum attacks. This thesis aims to design resource allocation algorithms to improve the energy efficiency, resiliency, and security of future optical networks, and the main contributions can be summarized into three parts as follows. (1) Proposal to Improve Energy Efficiency based on Novel Optical-Transmission Technologies: Recent advances in optical-transmission technologies bring new possibilities for enhancing energy efficiency for future optical networks. I have designed resource allocation algorithms for two emerging optical-transmission approaches, namely pluggable optics and power profile monitoring (PPM). Regarding pluggable optics, I have comprehensively investigated the energy efficiency of different ''IP over Wavelength Division Multiplexing'' (IPoWDM) network architectures with ZR/ZR+, and evaluated the advantages of ZR/ZR+ compared to long-haul muxponders. Specifically, I have proposed a pragmatic power consumption model for optical nodes and designed an auxiliary-graph-based approach to quantify the power consumption of different IPoWDM network architectures. Regarding PPM, I have proposed and investigated the novel optimized monitoring placement (OMP) problem for PPM to minimize the monitoring power consumption. Then, I have quantitatively compared the power consumption of PPM vs. the traditional monitoring technique, Optical Time-Domain Reflectometer (OTDR), with both an Integer Linear Programming (ILP) model for small-scale scenarios and an efficient heuristic algorithm for large-scale scenarios. These works provide guidelines for the deployment of ZR/ZR+ and PPM to improve the energy efficiency of future optical networks. (2) Proposal to Enhance Network Resiliency with Proactive and Reactive Approaches: A key challenge in network virtualization for future optical networks is to improve network resiliency against possible failures and enable swift recovery after failures. I have investigated proactive solutions to improve network resiliency against double-link failures and reactive solutions to provide swift network recovery under massive failures. Specifically, to provide network resiliency against any possible double-link failures in the network, I have investigated capacity sharing among existing virtual networks to reduce the additional resources to enhance network resiliency. More specifically, I have explored two strategies to enhance resiliency through capacity sharing, one employing a dedicated slice to improve resiliency and the other without such a slice. To solve this problem, I have formulated an ILP model and designed a scalable local-search-based heuristic algorithm. Regarding swift recovery, I have introduced a novel problem, called progressive slice recovery (PSR) with guaranteed reachability among network nodes (network connectivity) and reachability of data centers (content connectivity). Specifically, I have investigated to determine a recovery sequence for slices to swiftly recover the slices and, in the meantime, consider accelerating the provision of network services with connectivity constraints. To solve the PSR problem, I have developed an ILP model and a two-phase progressive slice recovery algorithm based on column generation and deterministic rounding. The numerical results confirmed that network resiliency can be guaranteed with less additional resources through proactive approaches and shorter recovery time using reactive approaches. (3) Proposal to Improve Network Security with Quantum Technologies: Quantum computers pose emerging security threats to future optical networks, making it critical to explore novel algorithms and architectures with quantum technologies. I have investigated novel resource allocation algorithms and architectures for QKD and quantum networks, enhancing their ability to securely distribute keys among nodes with guaranteed information-theoretic security. First, I have introduced a novel problem of Routing, Channel, Key-rate and Time-slot Assignment (RCKTA) to allocate quantum resources for QKD networks. This problem considers incorporating the advantages of trusted relays to increase the key rate and optical bypassing to reduce the unnecessary QKD modules (i.e., transceivers for QKD networks). To solve the RCKTA problem, I have devised a Mixed Integer Linear Programming (MILP) model and a near-optimal heuristic algorithm, achieving a substantial increase in acceptance ratio. Next, I have designed a novel QKD network architecture integrating Information-Centric Networking (ICN) to address the challenges of limited secret key rates and diverse service requirements. This architecture can accelerate key distribution with in-network key caching of ICN and meet diverse service requirements by analyzing the requirements based on semantic information. Lastly, I have investigated link configuration for fidelity-constrained routing and purification (LC-FCRP) in quantum networks, developing a MILP model and a shortest-path-based fidelity determination algorithm to optimize the fidelity of distributed entanglements. These approaches significantly increase the performance of QKD networks and quantum networks to guarantee the security of future optical networks.File | Dimensione | Formato | |
---|---|---|---|
2025_01_Zhang.pdf
accesso solo da BNCF e BNCR
Dimensione
8.3 MB
Formato
Adobe PDF
|
8.3 MB | Adobe PDF |
I documenti in UNITESI sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
https://hdl.handle.net/20.500.14242/207192
URN:NBN:IT:POLIMI-207192