Adapting the Machine Learning Grid Prediction Models for Forecasting of Resources on the Clouds

As the grid is loosely coupled congregation of geographically distributed and heterogeneous resources, the efficient scheduling of their resources require the support of a sound Performance Prediction System (PPS). The performance prediction of resources is helpful for both Resource Management Systems and users to make optimized resource usage decisions. In this paper, it is discussed about the ways of several Machine Learning (ML) PPSs that span over several grid resources in several dimensions. Taxonomy is used to categorize and identify approaches which are followed in the implementation of the existing PPSs for grids and for their adaptation in the clouds. A test framework is proposed to evaluate their capability of serving in the clouds environment.