Folksonomy-based reasoning for content dissemination in mobile settings

Modern mobile phones have become tools for the creation and consumption of digital media. There exist cases where the people producing such media content, and those interested in receiving it, tend to be living in the same geographical area. Delay Tolerant Networking (DTN) protocols have started to be investigated as an effective means to distribute content in these dynamically changing settings. The main challenge addressed by researchers so far has been the maximisation of delivery probability, while also minimising the overall network overhead (e.g., number of replicas in the system, messages' path length). Another important challenge that has received little attention so far is how to understand what content users are interested in receiving. The assumption often made is that users have a well defined and up-to-date profile describing their interests, and that content has been classified by means of a shared taxonomy. However, experience with the Web 2.0 demonstrates that, in most cases, neither assumptions hold. We thus propose a light-weight mechanism that dynamically learns what users are interested in based on the tags they use when they create and/or consume content. To maximise the chances of bringing relevant content to interested users, we apply a tag-expansion technique to enrich content descriptions beyond the folksonomy used by a single user to those spoken by the local community. We integrate this approach within a source-based DTN protocol we have previously developed, and evaluate its performance via simulation using real datasets.