Open source adaptive backup extensions for enterprise environment

This paper presents the procedure applied currently on a large network at Carnegie Mellon University to perform backup for the Linux computers, servers, and clusters, together with the software extensions needed to implement it. A study of user needs resulted in development of two backup scenarios, one of them using local disks and the other one using network storage. Starting from the existing backup software named TOB, the authors developed a series of extensions that adaptively change the configuration of TOB on the fly and handle the backup archives depending on the available storage, history, and file properties. The TOB configuration is created based on a set of prototypes and configuration files which can be the same for most of the computers and independent of the operating system flavor or version. To complete the usability of TOB, a web based GUI extension was developed to handle listing and recovery. The advantages of the solution presented here consist in more efficient storage management, more robust automated backup process, and minimal or no initial setup for the computers.