Collecting data about moving objects in a distributed environment

This paper presents and compares strategies for collecting data about moving objects in a distributed environment. Such an environment is composed of servers in charge of recording events that occur in particular partitions of a geographic space. This work shows that strategies based on the number of objects or the number of move in events are practical alternatives for collecting large amount of data about moving objects per time unit. Our study was performed in a simulated environment with different distributions of servers in the space, in particular, uniform, normal and cluster distributions. Among these distributions, collecting data on a uniform distribution of servers outperforms the collection of data on a normal or cluster distribution of servers.