Bachelor thesis/Master thesis

Title

PollenPile, Resource Discovery on Top of Instant Messaging Networks

Description

Resource discovery in dynamic environments is a tough task, whether this environment is a corporate information system or the World Wide Web.

The pattern of behavior followed by mobile software agents similar to the foraging pattern of honeybees provides a scalable method for discovery that consumes less time for a complete computation and less bandwidth.

Instant Messaging (IM) networks are virtualized P2P networks, and they can be utilized for information sharing. IM networks are already established, used by a very large population and easily deployable on client machines. Using such a Peer to Peer network instead of a centralized one (e.g. facebook, twitter …) provides maximum privacy to the user since the information are stored locally on her PC and access to it by visiting agents is totally controllable. This releases the user from struggeling with privacy policies and settings.

Semantic annotation opens new horizons; attributes like: "Suitable for children" or "MobileOK" have a unified meaning and can be used automatically by machines to filter out or recommend web content. POWDER, a new W3C recommendation, provides high flexibility in resource grouping and in the semantic level of annotation.

The agent runtime and connection through IM network are implemented already in previous work (PurpleBee framework). Expected results of proposed thesis include developing the Firefox add-on which will handle POWDER documents in two ways (read/write) in addition to the connection to a local data storage (read/write).

Ansprechpartner