可重构柔性网络下基于预测的网络资源重构算法

Forecast-based network resource reconfiguration algorithm for reconfigurable flexible networks

研究了可重构柔性网络环境下的网络资源重构,为解决可重构嵌入式网络的构建、消亡导致的底层网络资源使用不均衡的问题,提出了基于预测的网络资源重构算法(FNRA)。与原有算法使用的周期性时间间隔不同,该算法基于已知信息预测网络重构时间间隔,通过对网络资源分类和重构请求滞后策略来保证网络重构时机的全局最优。在网络重构代价和建网成功率两个方面将FNRA与路径迁移和路径准入算法(PMPA)和虚拟网络准入算法(VNA)Ⅱ算法进行了比较,仿真结果表明FNRA的重构代价比PMPA降低40%,比VNAⅡ降低67%;FNRA的构网成功率比PMPA提高50%,比VNAⅡ提高27%。

A study of network resorce reconfiguration for reconfigurable flexible networks was conducted, and a forecast- based network resource reconfiguration algorithm （FNRA）was proposed to solve the usage imbalance of substrate network resources brought by reconfigurable embedded networks＇ dynamical joining and leaving. The FNRA takes full advantage of known information to calculate the interval of reconfiguration and uses this interval to replace the periodic time span in previous algorithms. In order to make the reconfiguration time global optimal, it takes the fol- lowing two measures：resources are divided into critical resources and general resources, and different algorithms are used for key resources and common resources respectively;according to the structure, the limit number of reconfiguration requests is deduced. The FNRA algorithm having was compared by a simple hysteretic simulation with the VNAII, a virtual network admission algorithm, and the path migration and path admission algorithm （PMPA）. The results showed that the FNRA reduced the reconfiguration cost by 40% than that of the PMPA and 67% than that of the VNAII, and improved reconfigurable embedded networks＇ request reception rate by 50% than that of the PMPA and 27% than that of the VNAII.