JSON Based Web Services Virtualization Platform

Abstract

Delivering working software on time is a major challenge in the software industry. This challenge becomes more challengeable when the initial requirements want to be changed while the development is in progress. During the last few decades, different software development methodologies have been emerged by different groups to minimize those problems and improve the quality of the development process, hence, to improve the quality of the final product. Currently, the dimension of this problem has become more complex since the emergence of distributed software systems. Nowadays, most of the enterprise level applications are heterogeneous, consists of different service, application modules and communication patterns. And Web services are considered as the most efficient communication method among applications. In such an environment, there should be a well-organized engineering process to produce quality output. The parallel development process is considered as one of the best practice to improve efficiency by sharing the responsibilities of different components to different engineering teams through documentation and meetings. Though, communication of requirements changes is harder and takes more time. Sometime it may cause another team to hold their works until another team completes their tasks. This is one of the major drawbacks of the parallel development process. This dissertation proposes a solution to overcome this problem and improve the efficiency of the parallel development process through virtualizing web services using JSON. The proposed solution contains two major parts. First part handles virtualization of different services and the second part handles arbitrary data generation while processing the HTTP requests. The proposed solution uses JSON as the key format of persisting service definitions since it's ease of maintainability and accessibility over XML and traditional database schemas. Data generation part can process requests and produce complying formatted random data asynchronously for each request. Simple user interfaces allow users to work with the system without deeper knowledge about computer science. So technical and non-technical people like business analysts can interact with the system and communicate requirement changes rapidly to other stakeholders.