PROCESS-ORIENTED SERVICE COMPOSITION BASED ON OPENAPI SPECIFICATION (OAS) 3.0

• Main Author: Muhamad, Wardani
• Format: Dissertations
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/83745
• Institution: Institut Teknologi Bandung

The development of IT services, in accordance with the goals of service computing, must be able to bridge the gap between IT and business with the main technology that can be chosen to make this happen is web services. Of the various approaches or architectural styles available for developing web services, REST, or also more commonly known as web API, has become the choice most widely adopted by software developers in the last decade. As a piece of software, a web service provides an interface that service users can learn about by providing a service description. Service description is the main component and has an important role in designing a combination of various services through service composition. Web API, which is the architectural style most chosen by developers, uses the OpenAPI Specification (OAS) as a standard for presenting service descriptions. OAS documents provided by service providers become a knowledge base that service users can manage in their service repositories. To meet the development of increasingly complex business services, businesses must continue to strive to find effective and efficient ways by utilizing the collection of web services they own, and web services provided by other businesses. Service composition that combines several web services to fulfill increasingly complex business services is an option to answer these challenges. By implementing service composition, IT service development can be done quickly because not all business functions have to be built from scratch and it is more economical because you can use IT services that have been provided by other businesses. Designing a new service (composite service) through service composition requires a comprehensive understanding of both business and IT aspects. In service provider organizations, there are at least domain experts who master high-level business needs and software engineers who master software development in the form of web services. Both parties must be able to work together so that the service composition can work effectively. The existence of a knowledge gap between each can be an obstacle to the success of service composition. For this reason, in this research a conceptual model for composite service development is proposed which can be used as a reference for domain experts and software engineers in designing, modeling, and developing composite services. In addition, to improve the meaning (semantics) of service descriptions in accordance with OAS standards, improvements to OAS are iv proposed so that they can describe semantic services and make them into a collection (dataset) of semantic web services. As proof-of-concept, two prototypes were developed, namely a virtual classroom service system and a semantic web service translator. A virtual class service system was developed to prove that the conceptual model of composite service development can be implemented, while a semantic web service translator was developed to automatically convert OAS documents into semantic web services. The semantic web service translator prototype was tested by extracting 106 OAS documents obtained via APIs.guru in the period 05-10 May 2021. As a result, 14,213 RDF triples in turtle format were obtained as a semantic web service dataset which can be a knowledge base for service consumers when they want to find service candidates needed in the service composition process. In addition, the use of a database as middleware in the process of converting OAS documents into semantic web services is able to provide an average efficiency of 14.59% compared to not applying the reuse of the same data schema. The virtual class service system can be accessed via https://smartcampus.my.id and is used to organize training/bootcamps organized by the Cyber Security Center of the School of Electrical and Informatics Engineering, Bandung Institute of Technology, Research Alliance Ko+Lab, Faculty of Applied Sciences, Telkom University, and Solo Christian Technology University in the period 31 July 6 September 2023 involving 151 students and 6 teachers. To test user acceptance of the virtual classroom service system, technology acceptance measurements were carried out using the UTAUT framework. Measurements were carried out to prove the influence of constructs that have an impact on intentions (behavioral intentions/BI) and habits (use behavior/UB) using the virtual classroom service system in the long term. As a result, it was concluded that BI is determined by performance expectancy (PE) and technology integration (TI), while UB is influenced by behavioral intentions (BI) and TI.