The term “cloud computing” has become a catch-all phrase for a wide variety of computing approaches. This variety ranges from AWS-style resource provisioning for easier and cheaper deployment of classic application stacks, through cloud-specific application server stacks like Google’s AppEngine and 10gen, and on to on-demand applications (SaaS) such as Salesforce.com. No matter what approach one chooses, the familiar architectural ingredients like the RDBMS are getting stretched beyond the breaking point, as requirements for size, reliability, and geographic span pit these solutions against the limits of technology and physics.

This talk examines three approaches to application data storage “in the cloud,” and shows how these approaches are going to force developers to rethink how they design, deploy, and manage their cloud-scale applications.