一种基于经济模型的数据复制收入预测函数

A Novel Revenue Prediction Function for Economically-Effective File Replication

  • 摘要: 提出了新的基于齐普夫分布和二项式分布相结合的预测函数,同时考虑了文件的序列相关性和地理位置相关性(基于二项式和齐普夫分布)。实验数据显示该方法能有效地提高网格环境下,网格任务访问数据的效率和最大吞吐量,预测效果优于文献2中的预测函数。

     

    Abstract: In this paper, a new prediction function based on Zipf and Binomial distribution is proposed, which considers both the sequence relation and geographic sequence. The novel algorithm improves the efficiency of file access and the throughput of grid. It turns out that, over the performed tests, the novel algorithm predicts the simulated values better than the prediction function in Ref2.

     

/

返回文章
返回