Embedded Databases Krithi Ramamritham Indian Institute of Technology Bombay, India krithi@cse.iitb.ac.in

Abstract

Builders and users of embedded devices, from sensors to network switches, are demanding more and more from them, leading to an ever increasing range of functionalities and features being added to these devices. Many embedded applications need to deal with significant amounts of data which necessitates a local database management system. However, scaling down traditional database applications as well as DBMSs has not been very easy as many embedded devices are characterized by comparatively lesser memory and stable storage.

Motivated by the many issues that arise in the development of a lightweight relational DBMS for embedded devices, we have been developing a library of components that provide a toolbox for crafting a DBMS appropriate for an application given the constraints of its embedded execution environment. In the process, we have developed novel data storage and indexing techniques as well as several query processing techniques motivated by memory constrained environments. Our solution suite also considers issues like synchronization that arise when an embedded device shares data with a remote database.