对Pertel关于可适应寻径的批评的讨论

THE AMENDMENT FOR PERTEL'S CRITICISM OF ADAPTIVE ROUTING

本文针对Ｐｅｒｔｅｌ关于可适应寻径提出的批评［２］进行了深入的探讨，指出其结果的局限性，并进行了相应的改进．我们讨论了具有实际意义的缓冲器有限的情况，并把Ｐｅｒｔｅｌ的模型推广到对称性的Ｔｏｒｕｓ网络上，得出了低维ｋ－ａｒｙｎ－ｃｕｂｅ结构上更全面的结果．本文的工作不是要讨论ＡＲ或ＤＲ孰优孰劣，而着重于在实用的拓扑结构上，对更接近实际的问题进行定量的分析．文中给出了各种前提下相应的模拟结果，以供今后的工程实践参考．

This paper thoroughly discusses the criticism of adaptive routing proposed by Pertel. It finds some limitations of Pertel's work, and extends his model to a more practical one. The authors practically simulate the deterministic routing and the adaptive routing with finite buffers and a symmetry topology, theTorus, and get more general conclusions on the low dimensional k--ary n-cube. The work does.not intend to discuss which one is better, adaptive routing or deterministic routing,but to quantitatively analyze practical problems on existing topology. Results underdifferent conditions are given.