On the Cost-Benefit Tradeoffs of Cloud Storage Services for End Users and Service Providers

Abstract

Cloud storage is a data-intensive Internet service that synchronizes files from user devices, such as PCs and smartphones, with the cloud, offering means for users to easily backup data and perform collaborative work. This service has attracted a large interest from end users, industry and academia. Dropbox, which is one of leading companies in this market, report having more than 500 million registered users with 1 billion files upload daily. Despite the large popularity of cloud storage, little is known about the implications of typical user behavior patterns and policies adopted by popular services to the costs and benefits for end users and service providers. This thesis aims at investigating such cost-benefit tradeoffs for both parts jointly. Using data collected from Dropbox in various networks, we have developed new service policies, analytic models, and workload analysis tools to support providers and users in improving jointly their benefits in this kind of service.