چكيده لاتين
The rapid expansion of context-aware systems, driven by the advancement of ubiquitous computing and the Internet of Things, has intensified the need for seamless sharing of contextual information among heterogeneous systems. However, achieving interoperability in such environments remains a substantial challenge. Syntactic heterogeneity, including variations in data formats, data types, and communication protocols, as well as semantic heterogeneity, such as differing interpretations of concepts, measurement units, and data structures, often hinders systems from correctly interpreting and exchanging contextual information. As a result, developing middleware capable of addressing these discrepancies and ensuring reliable and efficient context sharing has become a complex and time-consuming endeavor.
To overcome these challenges, this dissertation introduces HarmonyCAS, a model-driven framework designed to streamline the development of publish–subscribe middleware and enhance interoperability in context-aware systems. HarmonyCAS incorporates a Domain-Specific Language (DSL) comprising three core components. The first component defines middleware elements, capturing the structure and behavior of the publish–subscribe system. The second component specifies context models that precisely represent contextual information and their interrelations. The third component provides mapping constructs responsible for handling syntactic and semantic differences across heterogeneous systems. Moreover, the framework integrates the concept of Ontology as a Service (OaaS) to enable a shared interpretation of concepts among systems, thereby strengthening semantic interoperability.
Experimental evaluations demonstrate that HarmonyCAS delivers stable and efficient performance. In experiments involving the transmission of 2,000 messages, no errors were observed, resulting in an error rate of zero. Under a load of 5,000 messages, the average response time was measured at only 49 milliseconds, indicating good scalability. Additionally, usability assessments revealed that user satisfaction scores, including ease of use, ease of learning, perceived usefulness, and overall satisfaction, exceeded 80 percent. These results indicate that HarmonyCAS effectively facilitates context sharing and meets essential interoperability requirements. Overall, the findings of this research suggest that HarmonyCAS can serve as both a scientific and practical approach for simplifying the development of publish–subscribe middleware and addressing interoperability challenges in context-aware systems, thereby paving the way for broader adoption in real-world environments.
