Caching With Meaning: Personality Web Caching Xingjun Zhang xjzhangxjtu'edu'cn Xian Jiaotong Univers PowerPoint PPT Presentation

presentation player overlay
1 / 20
About This Presentation
Transcript and Presenter's Notes

Title: Caching With Meaning: Personality Web Caching Xingjun Zhang xjzhangxjtu'edu'cn Xian Jiaotong Univers


1
Caching With Meaning Personality Web Caching
Xingjun Zhangxjzhang_at_xjtu.edu.cn Xian
Jiaotong University
2
Outline
  • Introduction
  • Generic Web caching scheme
  • Proxys personality
  • Personality proxy architecture
  • A novel document replacement algorithm
  • Personality Web caching Information Grid
  • Conclusion

3
Introduction
  • Performance is a major concern in the Web
  • The most popular applications on the Internet
  • Its size is growing in exponential order
  • Web caching is the most widely used method to
    improve Web performance
  • Advantages of Web caching
  • Reduces bandwidth usage
  • Lightens Web server loads
  • Lessens user-perceived latency

4
Generic Web caching scheme
  • A common form - caching at proxies
  • Proxies placement
  • close to the server
  • close to users
  • Proxies collaboration
  • Hierarchical
  • Distributed
  • Hybrid
  • Proxy management policies information
  • Operational information
  • Web documents properties

5
Proxys personality
  • User community
  • Users of the same Intranet, students of the same
    University, etc.
  • Different user communities have different
    business, so different interests
  • A proxy resides in users network edge and
    delegates the common interests of user community.
  • We call the community common interests as proxy
    personality

6
Characterize proxys personality
  • Grasping communitys interests need to understand
    the content accessed by the community
  • Understanding content need a semantic environment
  • Ontology-based personality characterizing

7
Ontology
  • An ontology formally defines a common set of
    terms that are used to describe and represent a
    domain
  • A domain refers to a specific subject, or a field
    of knowledge
  • A simple ontology can be implement using a
    keyword hierarchies

8
Cached documents view
  • Generic proxy view cached documents as a weakly
    linked set of resources
  • Ontology-based the cached documents and
    relationships among them can be explored

9
Personality proxy architecture
10
Advantages of Personality proxy
  • Optimize document replacement policy
  • Content-based replacement policy
  • Optimize the proxy collaboration
  • Exchange popular information
  • Implement new prefetching policies

11
Prototype
  • Construct ontology
  • Concept hierarchy tree
  • Documents semantic information (Meta-tag, title,
    link, keyword)
  • Classical directory (Open Directory Project)
  • Documents get a semantic weight
  • Problems
  • Chinese documents need word syncopating
  • Concept classify is a difficult problem

12
A novel replacement algorithm(1)
  • Existing document replacement algorithms

13
A novel replacement algorithm(2)
  • Web access pattern analysis
  • 10 Websites take on 80 user access
  • Same Website has different access scale in
    different proxy
  • Websites with equally access scale have different
    unique request in a proxy
  • A coarsely granular ontology
  • Regard Websites content attribute as an
    ontology
  • Give Website a role value in a proxy

14
A novel replacement algorithm(3)
  • Website role value
  • SRp Website role value, document p comes from
    this Website
  • Ds access scale to Website s
  • Us unique access scale to Website s
  • ? parameter to balance SRp value

15
A novel replacement algorithm(4)
  • Site-role based replacement algorithm
  • Replacement weight of document p

Boeing
NLANR
16
Grid
  • Computational Grids
  • basic layer for processing power by distributing
    massive computational tasks to numerous resources
    over matching communication links.

data archives
MPPs
clusters
workstations
wide area
instruments
17
More grids
  • Three-layer Grid architecture
  • Information and knowledge Grids
  • allowing access to dispersed information, and
    knowledge discovery and extraction from spread
    knowledge resources. They make use of cognitive
    techniques and tools such as data mining, machine
    learning, content semantics, ontology
    engineering, and intelligent agents

18
Personality Web Caching Information Grid
  • Web gives supports to Grid to some extent
  • Personality Web caching give supports to
    Information Grid
  • The important difference between the Web and the
    Grid is the collaboration ability. The
    personality Web caching, understanding the
    informations meaning and structuring data in
    certain degree, can make the proxies share
    collaborative information spaces and support
    synchronous collaboration.
  • Information Grids issues are relevant to the Web
    caching and CDN, such as homogeneous access to
    heterogeneous resources, distribution and
    delivery of information, content management,
    location of information, middleware for
    information access and integration

19
Conclusion
  • Using ontology-based method to characterize the
    proxy
  • Proposed a new personality proxy architecture
  • Proposed a novel contend-based document
    replacement algorithm
  • Personality Web caching supports Information Grid

20
  • Thanks!
Write a Comment
User Comments (0)
About PowerShow.com