Acta Metallurgica Sinica(English letters) ›› 2012, Vol. 19 ›› Issue (5): 83-91.doi: 10.1016/S1005-8885(11)60304-X

• Networks • 上一篇    下一篇

Storage capacity model for cloud download

徐嬴颖1,敖乃翔2,陈常嘉1,黄琰3   

  1. 1. 北京交通大学
    2. 北京交通大学电子信息工程学院
    3. 腾讯公司
  • 收稿日期:2012-04-05 修回日期:2012-06-28 出版日期:2012-10-31 发布日期:2012-10-08
  • 通讯作者: 徐嬴颖 E-mail:yingying_xu@tom.com
  • 基金资助:

    北京交通大学基本业务费“多点到多点传输机制建模及协议设计

Storage capacity model for cloud download

XU Ying-ying, AO Nai-xiang, CHEN Chang-jia   

  1. School of Electronics and Information Engineering, Beijing Jiaotong University, Beijing 100044, China
  • Received:2012-04-05 Revised:2012-06-28 Online:2012-10-31 Published:2012-10-08

摘要:

With the rise and world wide deployment of cloud utilities, the principle of the cloud download is proposed to provide high-quality file content distribution by using dedicated servers as cloud cache to guarantee the data availability and enhance the data transfer rate. As the system scales up to a large population, how to design appropriate storage capacity of cloud cache is a key challenge for cloud download. In this paper, primary elements impacting on storage capacity are explored through deliberating in large-scale commercial cloud download system, i.e. general user usage pattern and available period. And based on statistical analysis of real-world traces, we formulate storage capacity related to these two elements, which is the original contribution different from all previous works. This model gives guidance of potential system policy design. Finally, the effectivity of this model is demonstrated by simulation results compared with empirical data of practical system.

关键词:

cloud download, storage capacity model, available period, usage pattern

Abstract:

With the rise and world wide deployment of cloud utilities, the principle of the cloud download is proposed to provide high-quality file content distribution by using dedicated servers as cloud cache to guarantee the data availability and enhance the data transfer rate. As the system scales up to a large population, how to design appropriate storage capacity of cloud cache is a key challenge for cloud download. In this paper, primary elements impacting on storage capacity are explored through deliberating in large-scale commercial cloud download system, i.e. general user usage pattern and available period. And based on statistical analysis of real-world traces, we formulate storage capacity related to these two elements, which is the original contribution different from all previous works. This model gives guidance of potential system policy design. Finally, the effectivity of this model is demonstrated by simulation results compared with empirical data of practical system.

Key words:

cloud download, storage capacity model, available period, usage pattern