Unified Representation and Reuse of Federated Cloud Resources Configuration Knowledge

Abstract

The proliferation of tools for different aspects of cloud resource configuration processes encourages DevOps to design end-to-end and automated configuration processes that span across a selection of best-of-breed tools. But heterogeneities among configuration knowledge representation models of such tools pose vital limitations for acquisition, discovery and curation of configuration knowledge for federated cloud application and resource requirements. We propose an embryonic data-model for representing cloud resource configuration knowledge artifacts. We also propose a rule based recommender service, which empowers (1) incremental knowledge acquisition and curation, and (2) declarative context driven knowledge recommendation. The paper describes the concepts, techniques and current implementation of the proposed system. Experiments on 36 real-life cloud resources show efficient re-use of configuration knowledge by our approach compared to traditional techniques.