Making Data Storage Efficient in the Era of Cloud Computing

We enter the era of cloud computing in the last decade, as many paradigm shifts are happening on how people write and deploy applications. Despite the advancement of cloud computing, data storage abstractions have not evolved much, causing inefficiencies in performance, cost, and security. This dissertation proposes a novel approach to make data storage efficient in the era of cloud computing by building new storage abstractions and systems that bridge the gap between cloud computing and data storage and simplify development. We build four systems to address four data inefficiencies in cloud computing. The first system, Grandet, solves the data storage inefficiency caused by the paradigm shift from upfront provisioning to a variety of pay-as-you-go cloud services. Grandet is an extensible storage system that significantly reduces storage costs for web applications deployed in the cloud. Under the hood, it supports multiple heterogeneous stores and unifies them by placing each data object at the store deemed most economical. Our results show that Grandet reduces their costs by an average of 42.4%, and it is fast, scalable, and easy to use. The second system, Unic, solves the data in