Kachako: Towards a Data-centric Platform for Full Automation of Service Selection, Composition, Scalable Deployment and Evaluation

A service platform, ideally, should provide automation features for users to achieve their tasks by minimum human´s work. Although web services themselves could automate some of the human work, existing systems tend to lack such an automation point of view in their designs. Furthermore, scalability, availability, reusability and customizability should be provided altogether for a platform to be practically useful. We propose an architecture which ultimately automates everything in using services as much as possible. Exploiting the increasing number of freely available software, our Kachako platform transfers services themselves in a data-centric way to arbitrary servers that users wish to use. While Kachako is primary provided as a large integrated system with graphical user interfaces, Kachako is designed in a modular way to adapt various demands of users by allowing partial reuse of modules. End users can use Kachako without any programming work, while programmers can customize the system or add their own services very easily.