Automated object persistence for JavaScript

Abstract

Traditionally web applications have required an internet connection in order to work with data. Browsers have lacked any mechanisms to allow web applications to operate offline with a set of data to provide constant access to applications. Recently, through browser plug-ins such as Google Gears, browsers have gained the ability to persist data for offline use. However, until now it's been difficult for a web developer using these plug-ins to manage persisting data both locally for offline use and in the internet cloud due to: synchronization requirements, managing throughput and latency to the cloud, and making it work within the confines of a standards-compliant web browser. Historically in non-browser environments, programming language environments have offered automated object persistence to shield the developer from these complexities. In our research we have created a framework which introduces automated persistence of data objects for JavaScript utilizing the internet. Unlike traditional object persistence solutions, ours relies only on existing or forthcoming internet standards and does not rely upon specific runtime mechanisms such as OS or interpreter/compiler support. A new design was required in order to be suitable to the internet's unique characteristics of varying connection quality and a browser's specific restrictions. We validate our approach using benchmarks which show that our framework can handle thousands of data objects automatically, reducing the amount of work needed by developers to support offline Web applications. © 2010 International World Wide Web Conference Committee (IW3C2).