Data Management for mobile applications dependent on geo-located data

Abstract

An increasing number of mobile applications share location-dependent information, from collaborative applications and social networks to location-based games. In such applications, users are interested in information related to their immediate surroundings or destination when moving instead of data referring to events or state in distant areas. The current database systems enforce uniform consistency models that do not take into consideration data geographical locality, requiring applications to implement ad-hoc solutions that are sub-optimal at best, and can lead to poor performance in the worst case. In this paper, we argue in favour of consistency models where data location is a key property of data items that is leveraged to govern the operation of replication protocols and the guarantees provided to data accessed by users. To illustrate this, we present FocusDB, a new data management system designed to leverage both object and client location to combine stronger and weaker levels of consistency on a per-object basis. The system discussed here represents a first step in a larger ongoing research effort focused on deriving new consistency models and replication protocols that leverage our previous observation.