DHT网络中的多维复杂查询处理方法研究

摘要

DHT网络中的高级查询处理是关系其应用领域拓展的重要问题,也是学术界与工业界共同关注的研究热点.基于Kademlia协议提出一种DHT网络中的多维复杂查询处理方法,其索引结构考虑了用户的查询偏好,使同类数据的存储位置具有相关性,同时巧妙地利用了Kademlia路由表的特点与更新方法,避免了索引维护过产生额外的通信开销,并且通过多点存储、沿路缓存改善了系统的容错性与负载均衡性.分析和实验表明,该方法以O( logN)的路由跳数复杂度和较低的开销实现了资源的多维复杂查询.%Advanced query processing is a critical problem for the application of DHT networks. It has attracted much attention from both academic and industrial community. This paper presented a technique for multi-dimensional complex query processing based on Kademlia. It takes user's preference into consideration so that homogeneous data is relevantly indexed. Furthermore; the index maintenance brings no extra communication cost by piggybacking on routing table re covery; besides the advantages in resilience and load balance. The analysis and simulation results show that it imple ments multi-dimensional complex query processing with O(logN) query length and low cost.