基于跳跃因子模型的不规则分布蓝色任务法
Unevenly Distributed Blue Tasks Algorithms Based on Skip-Over Model
-
摘要: 跳跃因子模型是处理系统超载的一种有效方法 它能够降低系统负载 ,并使系统缓慢降级 Deeply Red算法是一种简单的基于跳跃因子模型的任务请求丢弃方法 ,然而由于Deeply Red算法未考虑任务特性和CPU资源利用率 ,其性能较差 为了克服Deeply Red的缺点 ,提出不规则分布蓝色任务法 :静态不规则分布蓝色任务法 (SUDB)和动态不规则分布蓝色任务法 (DUDB) SUDB和DUDB算法都是通过降低高优先级任务对低优先级任务的干扰时间 ,来提高任务集的可调度性 模拟结果表明 ,SUDB和DUDB算法性能要优于Deeply Red算法Abstract: Skip over model is an effective method to deal with overload It can reduce the overload of the system and can make the system degrade gracefully Deeply Red algorithm is a simple scheduling method which is based on the skip over model It considers neither the characteristics of tasks nor the CPU utilization Therefore its performance is very low In order to overcome its shortcomings, two algorithms are proposed: static unevenly distributed blue tasks (SUDB) and dynamic distributed blue tasks (DUDB) SUDB and DUDB improve their performance by reducing the interfering time among the higher priority tasks and the lower priority tasks The simulation results show that the performance of SUDB and DUDB are much higher than that of the Deeply Red
下载: