基于组播树的多粒度波带静态疏导算法

研究了波带交换中的静态业务疏导算法。波带交换可以有效地减少波长交换的端口数量,但是当波带粒度值取固定值时,波带的粒度难以取得合适值。波带的粒度大,有助于减少交换端口的数量,但是波带利用率低;波带粒度小,有助于提高波带利用率,但是交换端口的数量多。为此,提出了多粒度的波带取值方法。根据静态业务疏导与组播路由的相似性,提出了利用构造组播树解决静态疏导问题的方法。另外,为了减少波带与波长交换平面互联的端口数量,采用了同目的地的波带疏导策略,并针对这一疏导策略提出了一种新的波带疏导辅助图。仿真结果表明,相对于固定粒度的波带取值,可以有效地减少交换端口的数量,并提高波带利用率。

A Multi-Granularity Static Waveband Grooming Algorithm Based on Multicast Tree

The static waveband grooming algorithm is studied. Waveband switching can effectively decrease the ports of wavelength switching. But the granularity of waveband is difficult to be set when it is constant. When the waveband granularity is large, the switching ports can be decreased while the waveband utilization is low. When the waveband granularity is small, the waveband utilization is high while the switching ports can be increased. To solve this problem, a multi-granularity waveband scheme is proposed. The static traffic grooming is similar to multicast routing in some degree. So the multicast tree is used to solve the waveband grooming problem. In order to decrease the connection ports between the waveband switching plane and wavelength switching plane, the same destination grooming scheme is adopted. A new waveband auxiliary graph is proposed for this grooming scheme. The simulation result shows that the scheme proposed can decrease the waveband switching ports and improve the waveband utilization ratio effectively.