In the context of the Industrial Internet of Things (IIoT) and machine learning–assisted manufacturing, this thesis analyses the latency and performance of Linux-based systems, as Linux remains the dominant family of operating systems due to its performance and efficiency. The objective of this work is to document the construction of four different Linux architectures and to compare their behaviour under both ideal and stressed conditions, across varying RAM capacities, focusing on the four most relevant components of the hardware stack. These architectures derive from technologies popularised for their security and resource isolation: namely, virtualisation, containerisation, and hybrid configurations combining the two. Such approaches enable the allocation of hardware resources to separate processes to enhance system security and resilience. The collected data was analysed using both classical statistical methods (ANOVA) and contemporary techniques such as functional statistics and bootstrapping, allowing for deeper and more substantive conclusions. The results provide a comparative assessment and highlight the advantages and costs of adopting these technologies in industrial environments.
In un contesto di IoT Industriale (IIoT) e produzione assistita da Machine Learning, questa tesi analizza la latenza e la capacità di sistemi Linux, essendo la famiglia di sistemi operativi dominante per la loro velocità ed efficienza. Questa tesi ha l’obiettivo di documentare la costruzione di quattro tipi diversi di architetture Linux e confrontarle tra di loro, in contesti ideali per la loro prestazione e sotto stress, con capacità di RAM diverse e focalizzandosi sui quattro sistemi più importanti dell’hardware. Le architetture derivano da tecnologie popolarizzate per la loro sicurezza e divisione di risorse; queste sono virtualizzazione, containers e casi ibridi dei due, e consentono di assegnare l’hardware a processi diversi allo scopo di migliorare la sicurezza e la resilienza del sistema. I dati vengono analizzati con metodi classici (ANOVA) e metodi contemporanei di statistica funzionale e bootstrapping, rendendo più approfondite e di sostanza le conclusioni della tesi, che hanno scopo comparativo e di dare luce ai vantaggi e ai costi dell’uso di queste tecnologie in un contesto industriale.
Exploring data acquisition performance in production-ready IIoT appplications by leveraging edge devices in machine tools
VALMORI, DANIEL
2024/2025
Abstract
In the context of the Industrial Internet of Things (IIoT) and machine learning–assisted manufacturing, this thesis analyses the latency and performance of Linux-based systems, as Linux remains the dominant family of operating systems due to its performance and efficiency. The objective of this work is to document the construction of four different Linux architectures and to compare their behaviour under both ideal and stressed conditions, across varying RAM capacities, focusing on the four most relevant components of the hardware stack. These architectures derive from technologies popularised for their security and resource isolation: namely, virtualisation, containerisation, and hybrid configurations combining the two. Such approaches enable the allocation of hardware resources to separate processes to enhance system security and resilience. The collected data was analysed using both classical statistical methods (ANOVA) and contemporary techniques such as functional statistics and bootstrapping, allowing for deeper and more substantive conclusions. The results provide a comparative assessment and highlight the advantages and costs of adopting these technologies in industrial environments.| File | Dimensione | Formato | |
|---|---|---|---|
|
Classical_Format_Thesis (16).pdf
accessibile in internet per tutti
Descrizione: Tesi Completa
Dimensione
6.65 MB
Formato
Adobe PDF
|
6.65 MB | Adobe PDF | Visualizza/Apri |
I documenti in POLITesi sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
https://hdl.handle.net/10589/246338