Mathematics and Computer Science
Content Addressable Storage Provider in Linux
Vesselin Dimitrov
Spring 2003
Mathematics and Computer Science Department
Denison University
Advisor: Jessen Havill
Content Addressable Storage Provider in Linux [pdf]
Abstract: In a Content Addressable Storage (CAS) model, data files are retrieved by their content rather than by a user-defined name. A digest over the content of the file is computed for each file and this digest acts as an identifier for retrieving the file. We implemented a basic file system in the Linux operating system kernel based on the CAS model. The file system, implemented as a character driver for a virtual CAS device, supports basic CAS functionality and is based on the concept of layered driver architecture. This CAS driver will become a part of a larger CAS system that can be used as a peer-to-peer system, as well as for single instance storage and reliable backup of files.