الفهرس 
Programming Model for Cloud Computing
Table of ContentsOnly 14 pages are availabe for public view
 |  | from | 126 |  |  |
| | | from | 126 | | |
Abstract
The Internet of Things (IoT) and context-awareness IoT applications are recently
involved in various domains such as smart homes, traffic, and healthcare. While developing
context-awareness IoT applications, the main challenge is managing the massive amounts of data
and events to get relevant context information. Context-awareness services are frequently
applied in context-awareness IoT applications’ domains. Therefore, there is a challenge while
developing these services in a unified and a flexible way to fit different IoT application domains.
In the cloud service model, there is a lack of concern about delivering the context information as
a service. This is needed for managing and controlling data and events to deliver only the
relevant context information to cloud services’ users.
This thesis proposes a domain-independent programming model that facilitates
developing context-awareness services for various IoT application domains. The proposed
programming model enables an adapted execution of the IoT services while hiding the context
management details. The proposed programming model applies an ontology-based approach for
context modeling, and it applies a rule-based approach for context reasoning.
The proposed programming model aims to strengthen the IoT middleware architectural
requirements and service requirements by utilizing the semantic technology. The middleware
service requirements would improve both data and event management mechanisms. It would
enable middleware architectural requirements: context-awareness, adaptability, abstraction,
service-based, and semantic interoperability. We demonstrate these supported features by a
detailed comparative study with the other related works.
We demonstrate the proposed programming model support of various domains by
developing two experiments on two different domains. Therefore, the effectiveness of applying
the context-awareness services is evaluated by achieving an average F1-score of 0.82,
comparable with the average F1-score of 0.73 of current state-of-the-art methods. Moreover, the
incorporation of essential compound activities cannot be handled by previous works.