1 / 13

Web Caching: Locality of References Revisited

Web Caching: Locality of References Revisited. Foong, A.P.; Yu-Hen Hu; Heisey, D.M. Department of Electrical and computer Engineering, University of Wisconsin Conference on IEEE International Networks, 2000. (ICON 2000). Proceedings., 2000 Page(s): 81 – 86 Kun-Ming Tien. Outline.

blithe
Download Presentation

Web Caching: Locality of References Revisited

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. Web Caching: Locality of References Revisited Foong, A.P.; Yu-Hen Hu; Heisey, D.M. Department of Electrical and computer Engineering, University of WisconsinConference on IEEE International Networks, 2000. (ICON 2000).Proceedings., 2000 Page(s): 81 –86 Kun-Ming Tien

  2. Outline • 1.Introduction • 2.Locality of reference in web accesses • 3.The Logistic Regression Model • 4.Implementation • 5.Applying to Web Caching • 6.Future Work & Conclusion

  3. 1.Introduction • The effort in this paper: --Determine what constitutes web locality --Find a good method for studying locality (Logistic Regression model) --Propose cache strategies based on the results • Effective web cache strategies are based on more than one feature

  4. 2.Locality of reference in web accesses

  5. 3.The Logistic Regression Model • It is widely used by the medical community

  6. The Logistic Regression Model(cont.) • Coefficients can be estimated by a suitable method( learning phase) • LR probability can be calculated for other objects( predication phase) • Y=1 if the web object re-accessed at least once , in the WF accesses

  7. 4.Implementation • Temporal Locality --X1=SINCE --X2=BHITS • Functional Locality --X3=SIZE --X4=TYPE • Topical/Contextual Locality --X5=NUM_LINKS --X6=NUM_IMAGES --X7=NUM_KEYWORD • Spatial Locality(dependency graph) --Primary & Secondary Objects

  8. Implementation(cont.)

  9. Implementation(cont.) • WF=100 & WB=100, NL=1000 & NP=10000 • Some Observation --multiple dimensions of locality exits --different sites exhibit different types of locality --primary objects show strong temporal & spatial locality --secondary objects have less temporal locality but strong topical locality

  10. 5.Applying to Web Caching • Lifespan • LR-LIFESPAN(cost=lifespan) • LR-LSIZE(cost=lifespan*size)

  11. Applying to Web Caching(cont.)

  12. Applying to Web Caching(cont.)

  13. 6.Future Work & Conclusion • Prefetch (Dependency graph) --We can predict accesses of secondary objects based on their features • Topical localityfull page parsing & content classification (XML,XHTML,optional tag) • Complex relationships among Cooperating caches

More Related