Due to the continued advances in computing hardware and wireless communication, integrated low-power sensing devices are becoming available. Such devices, when Internet-worked, enable monitoring in many different contexts.
For example, applications are emerging that involve the tracking of the positions of a set of moving objects. Positioning technologies such as GPS and Galileo coupled with the wireless communication technologies enable such tracking. The moving objects may be trucks, buses, taxis, police cars, or simply mobile phone users that use different location-based services (tourist services, location-based games, intelligent transport services, etc.).
Such applications are characterized by the high rates of updates arriving from the sensors. Queries that query the current state of the monitored phenomena have to be processed efficiently on the newest state of the data. In addition, the queries may be continuous, i.e., their results should be continuously maintained up-to-date as the new data from the sensors arrive.
How to organize the queries and the data in main-memory is an interesting challenge. The goal is to optimize the cost of data updates and the cost of maintaining the queries. A number of interesting trade-offs could be investigated. The memory hierarchy consisting of the RAM and a number of levels of CPU caches should be taken into account. In addition to empirical experiments, precise analytical cost modeling of the different proposed methods could be performed.