Abstract
This paper proposes a taxonomy for distributed network caches based upon discovery, dissemination, and delivery methods. We analyze the match between taxonomy categories and Web workloads, and classify current Web caching projects. Next, we describe a cooperative cache design known as server-directed proxy sharing (SDP). In SDP, proxy servers find cached copies by looking in local metadata directories. This speeds discovery and separates it from site selection, allowing proxies to pick sites using various criteria. Metadata is propagated by lazy prefetching, which piggybacks information about popular or related objects onto returned documents. Analytical simulation indicates SDP substantially reduces server load and connection denials as compared to standard proxy server caching.
| Original language | English |
|---|---|
| Pages (from-to) | 286 |
| Number of pages | 1 |
| Journal | Proceedings of the Hawaii International Conference on System Sciences |
| State | Published - 1999 |
| Event | Proceedings of the 1999 32nd Annual Hawaii International Conference on System Sciences, HICSS-32 - Maui, HI, USA Duration: Jan 5 1999 → Jan 8 1999 |
Fingerprint
Dive into the research topics of 'Taxonomy and design analysis for distributed Web caching'. Together they form a unique fingerprint.Cite this
- APA
- Author
- BIBTEX
- Harvard
- Standard
- RIS
- Vancouver