Decentralized Location-based Services

In recent years, location-based services have gained a lot of importance due to the proliferation of position-aware devices and the ubiquitous availability of wireless Internet access. This wide-spread interest in location-based services as well as their distributed nature of location-dependent data generation and consumption has resulted in a plethora of approaches for decentralized location-based services. Decentralized location-based services, however, face the challenges of peer churn and asymmetric Internet connectivity, resulting in an unreliable search, non-persistent storage, and inefficient dissemination of location-based data.

To solve these problems a qualitative analysis of existing approaches for decentralized location-based services is conducted in order to identify suitable design decisions. Based on the gained insights, a novel overlay termed Geodemlia.KOM is developed. The overlay uses a non-hierarchical and semi-structured overlay topology, which makes the overlay less susceptible to churn and guarantees a reliable area search. Furthermore, its built-in replication mechanism ensures the persistence of location-based data even under high churn rates. Finally, its multi-source content dissemination approach allows for the efficient exchange of location-based data over asymmetric links in the Internet. To demonstrate the superiority of Geodemlia.KOM, an in-depth performance and cost analysis is conducted using discrete event-based simulations as well as testbed experiments.