Synchronizable Databases for the Web
Alexy Khrabrov, Sumeet Sobti · 2000
In this paper we introduce a simple abstraction called a synchronizable database, and discuss its potential for managing a diverse range of web and internet applications. We illustrate its utility in improving performance of several existing web applications, and also describe several novel ways of using this abstraction. We give a two-layer architecture for implementing a synchronizable database. The abstraction forming the bottom layer is referred to as a summarizable database. In addition to traditional database functionality, it provides a facility for efficiently computing a digest(hash) of the records within any specified key range. The top layer is the synchronization facility. It implements an elegant protocol for synchronizing two summarizable databases over a limited bandwidth link such as the internet. We describe the synchronization facility in full, and briefly discuss our prototype implementation of a summarizable database.