Computer Science

Permanent URI for this collection

https://hdl.handle.net/10530/366

Browse

Browsing Computer Science by Author "Chiponga, Kudzai"

Now showing 1 - 1 of 1
  • Thumbnail Image
    Item
    A Contracts-Based Model for managing web services evolution
    (University of Zululand, 2015) Chiponga, Kudzai; Adigun, M.O.; Tarwireyi, P.
    Service-based systems need to be designed in such a way that they are able to accommodate the volatility of the environment in which they operate. Failure to evolve service-based systems will result in service providers losing their competitive edge. Additionally, failure to evolve these systems properly will have far-reaching negative impacts on all stakeholders, especially if disruptions are allowed to occur. However, evolving service-based systems in a non-disruptive manner is still a challenge both to the research community and industry. Over recent years, many organisations have been adopting technology-based service solutions in what has become a technology-driven business environment, with web services at the forefront of being a business-enabler. This dissertation focusses on developing a contracts-based model for managing the evolution of web services in a manner that is consistent and transparent to business partners. The design science research methodology was used to architect a model that can alleviate the challenge of evolving service oriented systems. It was found that technical web service contracts can be leveraged upon to manage and maintain consumers while evolution is carried out. A contracts-based service proxy was developed as an instantiation of the model. Experiments demonstrated that this proxy maintained compatibility between evolving services and existing consumers. The model developed in this research presented a cost effective solution to managing the evolution of webservices while reusing the same computing resources and cutting down on the development time needed in evolving web services. Even though the proxy introduced processing overheads, the resultant loss in service throughput was negligible, especially when we take into consideration the amount of time and effort taken in evolving services manually.