Escribe una PREreview

For decades, database research has focused on optimizing performance under fixed resources. As more and more database applications move to the public cloud, we argue that it is time to make cost a first-class citizen when solving database optimization problems. In this paper, we introduce the concept of cost intelligence and envision the architecture of a cloud data warehouse designed for that. We investigate two critical challenges to achieving cost intelligence in an analytical system: automatic resource deployment and cost-oriented auto-tuning. We describe our system architecture with an emphasis on the components that are missing in today's cloud data warehouses. Each of these new components represents unique research opportunities in this much-needed research area.