一种网格计算环境下WDM网路由和波长分配算法

波长路由光网络为网格计算动态提供端到端的专用光路,以满足网格计算应用的数据密集性,提前预留是网格计算中一种重要资源占用形式。目前波长路由光网络的提前预留研究多是关注用户请求的弹性调度,以缓解资源碎片影响。从路由和波长分配(RWA)算法缓解资源碎片影响的角度,分析了提前预留中RWA算法的特殊性,在已有路由和波长分配问题上增加了时间向量因素;并提出了最少资源碎片聚集(L-RFC)算法,引入了时间向量上的资源碎片影响因子作为波长分配权衡因素,将资源碎片影响分散在不同波长上。仿真证明,该算法即降低了阻塞率和资源碎片率,又提高了资源利用率,性能明显优于已有的RWA算法。

A novel WDM network routing and wavelength assignment algorithm in lambda grids

Wavelength routing optical networks provide end-to-end specific lightpaths for grid computing dynamic to satisfy the data density of grid computing applications, and the advance reservation is an important form of resource occupation in grid computing. At present, most researches on advance reservation of wavelength routing optical network force on the elastic scheduling of user's requirements for reducing the influence of resource fragmentation. From the angle of reducing the influence of resource fragmentation by routing and wavelength assignment (RWA), the particularity of RWA algorithm in advance reservation is analyzed and the time vector is also considered in RWA problem. Furthermore, the least resource fragmentation concentration (L-RFC) algorithm is proposed and the resource fragmentation impact factor of time vector is introduced as wavelength assignment balance factor, which separates the influence of resource fragmentation on different wavelength. Simulation results show that this algorithm not only reduces the blocking probability and resource fragmentation probability but also raises the utilization rate of resources, which is much better than conventional heuristic RWA algorithms.