The rapid growth and ubiquitous presence of the Internet of Things has laid the foundation for applications in various domains ranging from smart cities, transport systems and smart homes to healthcare, agriculture and smart grids. The data produced from these applications are diverse in nature with a diverse set of storage and processing requirements based on storage space, scope of the processed data and processing capacity required. These requirements have entailed a move away from storage and processing of data locally on end-devices to that on the cloud platforms for the Internet of Things. This paradigm shift of moving data storage and processing to the cloud has resulted in a steep increase in the number of cloud platforms. These cloud platforms are diverse in terms of the protocols they use, the application frameworks they offer, the security measures they implement and beyond. However, there is a lack of standard or framework to compare these cloud platforms, when selecting a platform, for building an Internet of Things application. Studies in existing literature compare platforms based on the generic requirements of Internet of Things applications and not based on any real world use-case scenario. The work in this thesis aims at bridging this gap in the existing literature. The contribution of this thesis is divided into two parts. In the first part, we have studied more than 20 cloud services offering different kinds of services for the Internet of Things including Infrastructure-as-a-Service, Platform-as-a-Service and Software-as-a Service. We have defined a novel taxonomy to classify the features of cloud services offered based on the PaaS. The taxonomy is applicable not only to the cloud platforms that we have studied but to any cloud platform for the Internet of Things. We have discussed the taxonomy in detail to elaborate the diversity of the features of the cloud platforms and the tradeoffs involved while selecting a platform for a use case. Furthermore, we have selected 12 cloud services offering PaaS from the above cloud services and have studied each of them in detail based on the taxonomy and presented the trends that we have observed for these cloud platforms. In the second part of the thesis, we have selected four of the above platforms to study in detail for a real deployment based on monitoring the environment and structural health of an ancient underground temple in Circo Massimo, Rome, as a part of a project in collaboration with archaeologists from the University of Trieste. We have discussed the challenges involved in building an end-to-end solution in a constrained environment such as this based on cloud platforms and the Internet of Things. For each of the cloud platforms, we have deployed a solution to store data from the site, process the data and visualize the data in accordance to the requirements of the archaeologists. We have studied the deployment for each cloud platform in detail and compared them based on their ease-of-use, documentation and the extent to which the cloud platforms fulfilled the requirements for the use-case.
La rapida crescita e l'onnipresenza dell'Internet delle cose ha gettato le basi per applicazioni in diversi settori, dalle città intelligenti, ai sistemi di trasporto e case intelligenti, all’assistenza sanitaria, all’agricoltura. I dati prodotti da queste applicazioni sono di natura diversa con diverse esigenze di storage e di elaborazione. Tali requisiti hanno comportato un allontanamento dalla memorizzazione ed elaborazione locale, e uno spostamento sulle piattaforme cloud. Questo cambiamento di paradigma ha portato ad un forte aumento del numero di piattaforme cloud disponibili. Queste sono diverse in termini di protocolli che usano, i framework applicativi che offrono, le misure di sicurezza che implementano. Tuttavia, vi è una mancanza di standard o framework per confrontare tali piattaforme, quando si seleziona una piattaforma, per costruire un applicazione. Gli studi esistenti effettuano un confronto di piattaforme in base alle esigenze generiche di applicazioni e non sulla base di scenario di uso reale. Il lavoro in questa tesi si propone di colmare questa lacuna. Il contributo di questa tesi è divisa in due parti. Nella prima parte, abbiamo studiato più di 20 servizi cloud che offrono diversi tipi di servizi per l'Internet delle cose tra cui Infrastructure-as-a-Service, Platform-as-a-Service e Software-as-a Service. Abbiamo definito un nuova tassonomia per classificare le caratteristiche dei servizi cloud offerti sulla base del PaaS. La tassonomia è applicabile non solo alle piattaforme cloud che abbiamo studiato, ma a qualsiasi piattaforma cloud per l'Internet delle cose. Abbiamo discusso la tassonomia in dettaglio per elaborare la diversità delle caratteristiche delle piattaforme cloud ei compromessi coinvolti mentre la selezione di una piattaforma per un caso d'uso. Inoltre, abbiamo selezionato 12 servizi cloud offrono PaaS dai servizi cloud di cui sopra e abbiamo studiato ciascuno di essi in dettaglio sulla base della tassonomia e ha presentato le tendenze che abbiamo osservato per queste piattaforme cloud. Nella seconda parte della tesi, abbiamo selezionato quattro delle piattaforme sopra per studiare in dettaglio per una vera distribuzione basato sul monitoraggio dell'ambiente e della salute strutturale di un antico tempio sotterraneo a Circo Massimo, Roma, come parte di un progetto in collaborazione con archeologi dell'Università di Trieste. Abbiamo discusso le sfide coinvolti nella costruzione di una soluzione end-to-end in un ambiente vincolato come questo sulla base di piattaforme cloud e l'Internet delle cose. Per ciascuna delle piattaforme cloud, abbiamo implementato una soluzione per memorizzare i dati dal sito, elaborare i dati e visualizzare i dati in base alle esigenze degli archeologi. Abbiamo studiato la distribuzione di ogni piattaforma cloud in dettaglio e loro confronto in base alla loro facilità d'uso, la documentazione e la misura in cui le piattaforme cloud soddisfano i requisiti per il caso d'uso.
Cloud platforms for the Internet of Things : how do they stack up in a real world application ?
DOLUI, KOUSTABH
2016/2017
Abstract
The rapid growth and ubiquitous presence of the Internet of Things has laid the foundation for applications in various domains ranging from smart cities, transport systems and smart homes to healthcare, agriculture and smart grids. The data produced from these applications are diverse in nature with a diverse set of storage and processing requirements based on storage space, scope of the processed data and processing capacity required. These requirements have entailed a move away from storage and processing of data locally on end-devices to that on the cloud platforms for the Internet of Things. This paradigm shift of moving data storage and processing to the cloud has resulted in a steep increase in the number of cloud platforms. These cloud platforms are diverse in terms of the protocols they use, the application frameworks they offer, the security measures they implement and beyond. However, there is a lack of standard or framework to compare these cloud platforms, when selecting a platform, for building an Internet of Things application. Studies in existing literature compare platforms based on the generic requirements of Internet of Things applications and not based on any real world use-case scenario. The work in this thesis aims at bridging this gap in the existing literature. The contribution of this thesis is divided into two parts. In the first part, we have studied more than 20 cloud services offering different kinds of services for the Internet of Things including Infrastructure-as-a-Service, Platform-as-a-Service and Software-as-a Service. We have defined a novel taxonomy to classify the features of cloud services offered based on the PaaS. The taxonomy is applicable not only to the cloud platforms that we have studied but to any cloud platform for the Internet of Things. We have discussed the taxonomy in detail to elaborate the diversity of the features of the cloud platforms and the tradeoffs involved while selecting a platform for a use case. Furthermore, we have selected 12 cloud services offering PaaS from the above cloud services and have studied each of them in detail based on the taxonomy and presented the trends that we have observed for these cloud platforms. In the second part of the thesis, we have selected four of the above platforms to study in detail for a real deployment based on monitoring the environment and structural health of an ancient underground temple in Circo Massimo, Rome, as a part of a project in collaboration with archaeologists from the University of Trieste. We have discussed the challenges involved in building an end-to-end solution in a constrained environment such as this based on cloud platforms and the Internet of Things. For each of the cloud platforms, we have deployed a solution to store data from the site, process the data and visualize the data in accordance to the requirements of the archaeologists. We have studied the deployment for each cloud platform in detail and compared them based on their ease-of-use, documentation and the extent to which the cloud platforms fulfilled the requirements for the use-case.| File | Dimensione | Formato | |
|---|---|---|---|
|
2016_12_DOLUI.pdf
accessibile in internet per tutti
Descrizione: CLOUD PLATFORMS FOR THE INTERNET OF THINGS: HOW DO THEY STACK UP IN A REAL-WORLD APPLICATION?
Dimensione
1.44 MB
Formato
Adobe PDF
|
1.44 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/10589/131959