基于Bloom滤波器的对等网多关键字检索

Multi-keyword search over P2P based on Bloom filter

现有基于Bloom滤波器(BF)的对等网(P2P)检索,由于索引表的不断增长且不能确定数据量的上限,存在两个问题:一是难以确定BF向量长度;二是不能高效处理P2P多关键字Top-k查询。提出了一种基于关键词频率进行分块的分块Dynamic Bloom Filter(BDBF)以解决上述问题;并给出了相应的P2P多关键字Top-k查询模型,即当节点传送BF时先传送高频DBF,如不能满足Top-k查询则继续传送次高频的BF。实验分析发现,该结构更能适应数据量的连续增长,降低网络传输流量,并能高效处理多关键字检索中的Top-k查询问题。

In keyword search over Peer-to-Peer （ P2P） based on standard Bloom Filter （ BF） , it is difficult to estimate the maximum number of the data sets because they are increasing continuously; hence, two problems show up： it is difficult to determine the upper value of the length of BF vector, and it cannot handle the multi-keyword search efficiently. To solve these problems, the authors proposed a new structure called Block Dynamic Bloom Filter （ BDBF） which partitioned the keyword based on the frequency, and presented a new Top-k multi-keyword search model： the node sends the higher frequency DBF firstly, and then sends the secondary higher frequency DBF if need. The experimental results show that the proposed method can be applicable for the increasing data of index list, and it can also decrease the network＇s traffic and efficiently resolve the problem of Top-k query in multi-keyword search over P2P.