world wide web caching cs457 seminar yutao zhong 11/13/2001

15
World Wide Web Caching CS457 Seminar Yutao Zhong 11/13/2001

Upload: charlene-day

Post on 04-Jan-2016

215 views

Category:

Documents


1 download

TRANSCRIPT

Page 1: World Wide Web Caching CS457 Seminar Yutao Zhong 11/13/2001

World Wide Web Caching

CS457 Seminar

Yutao Zhong

11/13/2001

Page 2: World Wide Web Caching CS457 Seminar Yutao Zhong 11/13/2001

Problems

Network congestion Server overloading User-perceived latency

Page 3: World Wide Web Caching CS457 Seminar Yutao Zhong 11/13/2001

Effective solution: Web caching

Storing popular objects at locations close to the clients

client

Web serverWeb cache

Different deployments client-oriented, server-oriented, proxy

Page 4: World Wide Web Caching CS457 Seminar Yutao Zhong 11/13/2001

Advantages of Web caching

Reduces bandwidth usage Lightens Web server loads Lessens user-perceived latency

Page 5: World Wide Web Caching CS457 Seminar Yutao Zhong 11/13/2001

Proxy cache (I)

Standalone cache

client

client

client

cache

router

Web

Page 6: World Wide Web Caching CS457 Seminar Yutao Zhong 11/13/2001

Transparent cache Router level

cache

Web

client

client

client

router

cache

cache

Proxy cache (II)

Page 7: World Wide Web Caching CS457 Seminar Yutao Zhong 11/13/2001

Transparent cache Switch level

Proxy cache (III)

Web

client

client

client

L4 switch

cache

cache

cache

router

Page 8: World Wide Web Caching CS457 Seminar Yutao Zhong 11/13/2001

Design issues

Latency caused by cache miss (cache replacement, prefetching,…)

Stale data (cache coherence) Dynamic data (dynamic cache)

Page 9: World Wide Web Caching CS457 Seminar Yutao Zhong 11/13/2001

Cache replacement Traditional policies

LRU, LFU, FIFO Key-based policies

LRU-MIN, LRU-Threshold, HyperG Function-based policies

Bolot/Hoschka, LNC-R (least normalized cost-replacement)

Page 10: World Wide Web Caching CS457 Seminar Yutao Zhong 11/13/2001

Cache prefetching Prediction

server-based vs. user-based Scenarios

poll vs. push Target

content vs. means preresolving, preconnecting,

prewarming

Page 11: World Wide Web Caching CS457 Seminar Yutao Zhong 11/13/2001

Cache coherence: deals with stale data

Data validation HTTP assistance

Conditional GET (IF-Modified-Since) Date: date Last-Modified: date Expires: date Pragma: no-cache Cache-control: max-age

Page 12: World Wide Web Caching CS457 Seminar Yutao Zhong 11/13/2001

Client polling Server invalidation Adaptive TTL Piggyback invalidation

Cache coherence:implementing mechanisms

Page 13: World Wide Web Caching CS457 Seminar Yutao Zhong 11/13/2001

Dynamic data caching

Active caching cache computation CPU cost

Web server accelerator

Page 14: World Wide Web Caching CS457 Seminar Yutao Zhong 11/13/2001

Typical systems

CERN httpd: 1994

Harvest (Squid): 1995

Akamai

Page 15: World Wide Web Caching CS457 Seminar Yutao Zhong 11/13/2001

Summary Effective solution to Internet

scalability problem Issues and techniques Open problems

cache routing, dynamic data caching, security...