基于索引布隆过滤器的DDS自动发现算法

DDS automatic discovery algorithm based on index BF

数据分发服务(DDS)被广泛用于分布式系统的网络搭建,其中自动发现机制是DDS的关键部分。现有的DDS自动发现机制大都采用简单发现协议(SDP),但这种协议在大规模分布式系统的网络环境中会产生网络负载过高、匹配效率低下等问题。针对这些问题,提出一种基于索引布隆过滤器的轻量级DDS自动发现算法。该算法基于多维向量结构和索引值间的位操作设计一种索引布隆过滤器,用于压缩分布式系统网络节点间的传输信息,同时提供比标准布隆过滤器更低的误判率。结合索引布隆过滤器与SDP,能够减少DDS自动发现过程中的资源消耗并提高匹配效率。实验结果表明,在节点匹配率为10%的情况下,所提出的DDS自动发现算法相比基于标准布隆过滤器的SDPBloom算法,发现过程的数据包数量减少了46.39%,匹配时间缩短了73.30%。

Data distribution server(DDS)is extensively utilized in the network construction of distributed systems,and the automatic discovery mechanism is a key part of DDS.The simple discovery protocol(SDP)is used in most of the existing DDS automatic discovery mechanisms,which can cause problems such as high network load and low matching efficiency in large-scale distributed system network environment.On this basis,a lightweight DDS automatic discovery algorithm based on index bloom filter(BF)is proposed.Based on the multi-dimensional vector structure and the bit operation between index values,an index BF is designed to compress the information transmitted between nodes in the distributed system network and provide a lower misjudgment rate than the standard BF.The combination of index BF and SDP can reduce the resource consumption in DDS automatic discovery process and improve the matching efficiency.The experimental results demonstrate that when the node matching rate is 10%,the proposed DDS automatic discovery algorithm significantly can reduce the number of data packets by 46.39%and shorten the matching time by 73.30%compared with the SDPBloom algorithm based on the standard BF.