This thesis details the research activities on the investigations and experimentation performed for the development of a Software-Defined Networks and Network Function Virtualization based large-scale distributed hierarchical architecture for Software-Defined Networks. In the first step, this research started off with analyzing the state of the art enabling technologies by performing a qualitative cross-comparison among the available technologies (including container engines, orchestrators, and many other supporting tools) using set of Key Performance Indicators (KPIs). Then, in the second step, to understand the interplay of these technologies in virtualized service infrastructures, by considering the main outcomes of the qualitative analysis, the set of technologies including Docker as the container engine, Docker Swarm and Kubernetes as orchestrators with load balancing and service discovery capabilities are deployed on bare-metal and OpenStack cloud platforms. Experimental tests are conducted to identify the most suitable set of technologies in the high-load and industrial use-case of a smart farm using KPIs e.g., CPU utilization, memory footprint, network load, connection delay, and request completion time. In the third step, an innovative, distributive, and hierarchical framework based on the SDN paradigm is developed, which comprises two levels of SDN controllers to configure and monitor large-scale optical switches and networking functionalities. On top of that, Virtual Network Functions (VNFs) are optimally deployed and managed by a centralized orchestrator. The final step proceeds towards the formation of a novel methodology for the dynamic and reactive management of forwarding rules in a (potentially large-scale) SDN-based network, based on the knowledge of network topology, the power consumption of optical switches, the expected volume of traffic, and the variability of the actual traffic load.
Questa tesi descrive in dettaglio le attività di ricerca sulle indagini e le sperimentazioni eseguite per lo sviluppo di un'architettura gerarchica distribuita su larga scala basata su Software-Defined Networks e Network Function Virtualization per Software-Defined Networks. Nella prima fase, questa ricerca è iniziata con l'analisi dello stato dell'arte delle tecnologie abilitanti eseguendo un confronto incrociato qualitativo tra le tecnologie disponibili (inclusi motori di container, orchestratori e molti altri strumenti di supporto) utilizzando una serie di indicatori di prestazione chiave (KPI). ). Quindi, nella seconda fase, per comprendere l'interazione di queste tecnologie nelle infrastrutture di servizio virtualizzate, considerando i principali risultati dell'analisi qualitativa, l'insieme di tecnologie tra cui Docker come motore di container, Docker Swarm e Kubernetes come orchestratori con bilanciamento del carico e le funzionalità di rilevamento dei servizi sono implementate su piattaforme cloud bare-metal e OpenStack. Vengono condotti test sperimentali per identificare il set di tecnologie più adatto nel caso d'uso industriale e ad alto carico di una smart farm utilizzando KPI, ad esempio utilizzo della CPU, footprint della memoria, carico di rete, ritardo di connessione e tempo di completamento della richiesta. Nella terza fase, viene sviluppato un framework innovativo, distributivo e gerarchico basato sul paradigma SDN, che comprende due livelli di controller SDN per configurare e monitorare switch ottici su larga scala e funzionalità di rete. Inoltre, le funzioni di rete virtuale (VNF) vengono distribuite e gestite in modo ottimale da un orchestratore centralizzato. Il passo finale procede verso la formazione di una nuova metodologia per la gestione dinamica e reattiva delle regole di inoltro in una rete basata su SDN (potenzialmente su larga scala), basata sulla conoscenza della topologia di rete, il consumo di energia degli switch ottici, il previsto volume di traffico e la variabilità del carico di traffico effettivo.
Enabling Technologies and Hierarchical Control Plane Management of Software-Defined Networks / Shah, Awais Aziz. - ELETTRONICO. - (2022). [10.60576/poliba/iris/shah-awais-aziz_phd2022]
Enabling Technologies and Hierarchical Control Plane Management of Software-Defined Networks
Shah, Awais Aziz
2022-01-01
Abstract
This thesis details the research activities on the investigations and experimentation performed for the development of a Software-Defined Networks and Network Function Virtualization based large-scale distributed hierarchical architecture for Software-Defined Networks. In the first step, this research started off with analyzing the state of the art enabling technologies by performing a qualitative cross-comparison among the available technologies (including container engines, orchestrators, and many other supporting tools) using set of Key Performance Indicators (KPIs). Then, in the second step, to understand the interplay of these technologies in virtualized service infrastructures, by considering the main outcomes of the qualitative analysis, the set of technologies including Docker as the container engine, Docker Swarm and Kubernetes as orchestrators with load balancing and service discovery capabilities are deployed on bare-metal and OpenStack cloud platforms. Experimental tests are conducted to identify the most suitable set of technologies in the high-load and industrial use-case of a smart farm using KPIs e.g., CPU utilization, memory footprint, network load, connection delay, and request completion time. In the third step, an innovative, distributive, and hierarchical framework based on the SDN paradigm is developed, which comprises two levels of SDN controllers to configure and monitor large-scale optical switches and networking functionalities. On top of that, Virtual Network Functions (VNFs) are optimally deployed and managed by a centralized orchestrator. The final step proceeds towards the formation of a novel methodology for the dynamic and reactive management of forwarding rules in a (potentially large-scale) SDN-based network, based on the knowledge of network topology, the power consumption of optical switches, the expected volume of traffic, and the variability of the actual traffic load.File | Dimensione | Formato | |
---|---|---|---|
34 ciclo-SHAH Awais Aziz.pdf
accesso aperto
Descrizione: PhD Thesis
Tipologia:
Tesi di dottorato
Licenza:
Creative commons
Dimensione
8.04 MB
Formato
Adobe PDF
|
8.04 MB | Adobe PDF | Visualizza/Apri |
I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.