Abstract: Packet classification that classifies the incoming packets according to theirs head information is a key technology in network applications. Performance of packet classification algorithm is an important factor of delay and throughput of network. Tuple Folded Vector algorithm (TFV), introduced in this paper, is applied to multiple fields. After making a specification on theory and realization of this algorithm, it compares TFV with Tuple Space Search algorithm and TSBP on performance and shows its high efficiency.

Key words: Packet classification, tuple, tuple space, folded vector

摘要： 包分类是根据到达的数据包的包头信息将包按照一定的规则进行分类的过程，它是网络应用中的关键技术。包分类算法的性能对网络的时延和吞吐量有决定性的影响。本文基于元组空间提出了一种适用于多维大规则库的包分类算法：元组向量折叠算法。论述了该算法的原理和实现算法，并将其与基本元组空间算法和基于元组的位并行包分类算法做比较，阐述了它的优越性。

关键词: 包分类, 元组, 元组空间, 向量折叠