Lightweight Adaptive Backup System

Our goal is to create open source software for performing backups on a network of dedicated storage servers or on local disks. The design objectives were the following:

• minimalist initial configuration, low maintenance overhead, optimized resource requirements;
• frequency, type, contents, and storage location of the backup volumes is determined automatically and dynamically by the software, based on optimization algorithms or heuristic rules applied to the existing resources, so the available storage capacity is used in the most efficient way;
• design of the software is modular and flexible, with clear and simple APIs, so it can be easily extended or modified by end users.

To insure production quality for every release, this software will be actively used for backups on a medium scale network of Linux systems (servers, reserch clusters, and desktops), as well as on individual computers (laptops, home computers).

This academic project was initiated at Carnegie Mellon University. The software will be released under the LGPL.

Please see the Labs paradigm for a description of the ideas behind the design of this project.