A Service-Oriented Framework for GNU Octave-Based Performance Prediction

Cloud/Grid environments are characterized by a diverse set of technologies used for communication, execution and management. Service Providers, in this context, need to be equipped with an automated process in order to optimize service provisioning through advanced performance prediction methods. Furthermore, existing software solutions such as GNU Octave offer a wide range of possibilities for implementing these methods. However, their automated use as services in the distributed computing paradigm includes a number of challenges from a design and implementation point of view. In this paper, a loosely coupled service-oriented implementation is presented, for taking advantage of software like Octave in the process of creating and using prediction models during the service lifecycle of a SOI. In this framework, every method is applied as an Octave script in a plug-in fashion. The design and implementation of the approach is validated through a case study application which involves the transcoding of raw video to MPEG4.