摘要

Current Internet Protocol routers only support equal cost multi-path routing, which performs the random path selection or the traffic uniform distribution among equal-cost paths. In biology, an adaptive attractor selection model is presented to simulate the concentration changes of two kinds of Escherichia coli's mRNA in changing nutrition environments with bistability equations. Inspired by the metabolism behaviors of E. coli, we propose an adaptive path selection scheme Open Shortest Path First-path selection by attractor selection to dynamically select the transmission path by the real-time path quality. Here, the mRNA concentration is analogous to the path quality. Then, to reflect the multipath quality, multi-stability equations are adopted and redesigned. Our scheme consists of two main features. The first one is a redefined path-activity to indicate multipath transmission goodness, which is inversely proportional to the offset between current path quality and best path quality. And the second one is a new attractor expression of the multi-stability equations to concretely specify the effect of a stochastic item noise in the equations on the path selection. Compared with the greedy selection and the uniform random selection in file transfer protocol (FTP) service, our scheme gains better performance on reducing file transmission time, traffic throughput, and traffic dropped.