基于缓存回收的成本节约云服务算法研究

Cost saving cloud service algorithm based on cache recovery

为了解决现有的数据缓存回收算法不适用于混合云计算和大数据共享的问题,提出了一种新的基于缓存回收的云服务算法。该方法首先构建了一个混合云部署框架,包含预报和云缓存回收两个基本模块,根据统一的成本模型和改进的MLP自适应调整将窗口大小参数化,按照窗口参数或必需的缓存空间选择出一组在缓存中最近最少使用(LRU)的对象。仿真结果显示,相比三种典型的缓存替换算法GD-Size、LRU和LFU,提出的云服务算法每年能够节约成本11.43%,同时提高了转发字节命中率、延迟节约率和缓存命中率10%以上。

In order to solve the problem that the data cache recovery algorithms was not applicable to hybrid cloud computing and big data sharing,this paper proposed a new cloud service algorithm based on cache recovery. The method established a hybrid cloud deployment framework included prediction and cloud cache recovery module. According to the cost model based on the unified,it made the window size parameterized based on MLP self-adjustment. According to the window parameter or necessary cache space,it selected a set of least recently used（ LRU） object in the cache. The simulation results show that,compared with the traditional GD-Size,LRU and LFU of the typical cache replacement algorithms,the deployment of a cloud service algorithm can save data output cost of 11. 43% a year,at the same time improve the forwarding byte hit rate and delay saving rate and cache hit ratio above 10%.