Provenance in the Cloud:Why and How?

Abstract—Provenance is an important aspect in the ver-ification, audit trails, reproducibility, privacy and security, trust, and reliability in many fields ranging from art, food production, medical sciences, in-silico experiments, and dis-tributed computing. On the other hand, Cloud computing is the business model of distributed computing and is considered the next generation of computing and storage platforms. Cloud computing requires an extension of the architecture of distributed and parallel systems by using virtualization techniques. Key to this extensible architecture is to support properties such as compute “on demand ” and “pay as you go” model. Clouds are in use since a few years and they already ex-panded in the business domain (Amazon EC2, Microsoft Azure, IBM SmartCloud) and research environments (EUCALYPTUS, OpenNebula, Nimbus). Many research domains have already adopted Cloud technology into their existing computational and storage platforms and, thus, a shift of technology is in progress. In this paper, we present provenance description in comput-ing sciences. Then, we give an overview of Cloud architecture and answer why provenance is important for Cloud computing. We introduce a mechanism to include provenance in the Cloud which requires minimal knowledge and understanding of underlying services and architecture. Therefore, we detail the importance along with the characteristics identified and present a framework for provenance in Cloud computing. We assure trust by augmenting a Cloud infrastructure with provenance collection in a structured way and present first performance results of the extended architecture. Finally, we discuss the results and summarize challenges and open issues of provenance in Clouds. Keywords-provenance; research or open Clouds. I