Abstract: This paper proposes a keywords decompound Hash algorithm for quick flow classification based on keywords decompound. It receives the key by decompounding the keywords to small blocks, xored with selected primes and multiplied. Compared D-hash with IPSX, CRC32 and Bob Jenkins Hash algorithms. Experimental results indicate that the algorithm has good capability in rate, well-proportioned keys and collision rate. The classification rate is 18 Mp/s, D-hash can satisfy the quick net flow measurement.

Key words: flow classification, Hash algorithm, keywords decompound

摘要： 在Hash算法的基础上，提出一种基于关键字分解的快速流分类算法。通过把关键字进行分解，与预先选定的随机质数向量进行异或后相乘得出Hash值。实验结果表明，与IPSX、CRC32和Bob-Jenkins Hash算法相比，该算法在哈希值的计算速率、值的分布均匀性和冲突率等方面具有较好的性质，分类速度达到18 Mp/s，能满足快速网络流量测量的需求。

关键词: 流分类, Hash算法, 关键字分解

CLC Number: 

• TP393