Retrivability of data in ad-hoc backup

Abstract

This master thesis looks at aspects with backup of data and restore in ad-hoc networks. Ad-hoc networks are networks made between arbitrary nodes without any form of infrastructure or central control. Backup in such environments would have to rely on other nodes to keep backups. The key problem is knowing whom to trust. Backup in ad-hoc network is meant to be a method to offer extra security to data that is created outside of a controlled environment. The most important aspects of backup are the ability to retrieve data after it is lost from the original device. In this project an ad-hoc network is simulated, to measure how much of the data can be retrieved as a function of the size of the network. The distance to the data and how many of the distributed copies are available is measured. The network is simulated using User-mode Linux and the centrality and connectivity of the simulated network is measured. Finding the device that keeps your data when a restoration is needed can be like looking for a needle in a haystack. A simple solution to this is to not only rely on the ad-hoc network but also make it possible for devices that keep backups to upload data to others or back to a host that is available to the source itself.