Forecasted Popularity Based Lazy Caching Strategy 


Vol. 10,  No. 3, pp. 261-268, Aug.  2003
10.3745/KIPSTA.2003.10.3.261


PDF
  Abstract

In this paper, we propose a new caching strategy for web servers. The proposed strategy collects only the statistics of the requested file, for example the popularity, when a request arrives. At a point of time, only files with higher forecasted popularity are cached all together. Forecasted popularity based lazy caching (FPLC) strategy uses exponential smoothing method for forecast popularity of web files. And, FPLC strategy shows that the cache hit ratio and the cache transfer ratio are better than those produced by other caching strategy. Furthermore, the experiment that is performed with real log files built from web servers shows our study on forecast method for popularity of web files improves cache efficiency.

  Statistics


  Cite this article

[IEEE Style]

P. Cheol and Y. H. Yeong, "Forecasted Popularity Based Lazy Caching Strategy," The KIPS Transactions:PartA, vol. 10, no. 3, pp. 261-268, 2003. DOI: 10.3745/KIPSTA.2003.10.3.261.

[ACM Style]

Park Cheol and Yu Hae Yeong. 2003. Forecasted Popularity Based Lazy Caching Strategy. The KIPS Transactions:PartA, 10, 3, (2003), 261-268. DOI: 10.3745/KIPSTA.2003.10.3.261.