A comparative study of awareness methods for peer-to-peer distributed virtual environments

摘要

The increasing popularity of multi-player online games is leading to the widespread use of large-scale Distributed Virtual Environments (DVEs) nowadays. In these systems, peer-to-peer (P2P) architectures have been proposed as an efficient and scalable solution for supporting massively multi-player applications. However, the main challenge for P2P architectures consists of providing each avatar with updated information about which other avatars are its neighbors. This problem is known as the awareness problem. In this paper, we propose a comparative study of the performance provided by those awareness methods that are supposed to fully solve the awareness problem. This study is performed using well-known performance metrics in distributed systems. Moreover, while the evaluations shown in the literature are performed by executing P2P simulations on a single (sequential) computer, this paper evaluates the performance of the considered methods on actually distributed systems. The evaluation results show that only a single method actually provides full awareness to avatars. This method also provides the best performance results.