The growth of ubiquitous and pervasive computing systems with complex, heterogeneous, dynamic and mobile components, makes the utilization of structured overlay networks more desirable due to their inherent flexibility and extensibility to support modern day applications. However, the performance and dependability of structured overlay networks depend on their fault tolerance and self-managing capability to ensure an acceptable level of service in the face of mobility and anticipated churn. In this respect, this paper provides a systematic reference for the research work on the overlay geometries and churn handling mechanisms proposed in the literature to maintain the connectivity and reachability of overlay nodes in the presence of dynamic participants. Through an extensive literature search, we provide a taxonomic analysis of existing churn handling mechanisms, their qualitative comparison along with the detailed description of the factors behind the churn behaviour of the overlay participants in the existing applications. In addition, it highlights the key research issues to guide the future research in the field of churn handling.

具有复杂、异构、动态和移动组件的无处不在和普遍的计算系统的增长，使得结构化覆盖网络的利用更加可取，因为它们具有支持现代应用程序的固有灵活性和可扩展性。然而，结构化覆盖网络的性能和可靠性取决于它们的容错性和自我管理能力，以确保在面对移动性和预期流失时提供可接受的服务水平。在这方面，本文为文献中提出的覆盖几何和流失处理机制的研究工作提供了系统的参考，以在动态参与者存在的情况下保持覆盖节点的连通性和可达性。通过广泛的文献检索，我们提供了对现有流失处理机制的分类分析，它们的定性比较以及对现有应用程序中覆盖参与者流失行为背后因素的详细描述。此外，它突出了关键的研究问题，以指导未来在客户流失处理领域的研究。