John Heidemann / Papers / Multi-way Rendezvous in the Web

Multi-way Rendezvous in the Web
John Heidemann and Wenhui Zhao
USC/Information Sciences Institute


John Heidemann and Wenhui Zhao. Multi-way Rendezvous in the Web. Technical Report 98-687. University of Southern California Computer Science Department. [PDF] [alt PDF]


\emphWeb Rendezvous is the meeting of information publisher and subscriber, allowing the subscriber to create and maintain a list of links (indices) to publishers’ pages. Rendezvous is difficult because pages may move and unauthorized publishers may seek to add their pages to the index (spamming). Current rendezvous approaches are work-intensive or restricted to very limited domains. We provide a taxonomy of web rendezvous methods. We identify automated methods which allow a subscriber to maintain an index of many pages with only a fixed amount of effort. Our methods leverage existing web search engines to be robust to page movement, and use encryption to prevent index spamming. We have designed and implemented two different rendezvous methods; we summarize our experience using them, finding that we can securely track index changes in a few days with publisher support, and eventually even without publisher support.

Bibtex Citation

  author = {Heidemann, John and Zhao, Wenhui},
  title = {Multi-way Rendezvous in the Web},
  institution = {University of Southern California Computer Science Department},
  year = {1998},
  sortdate = {1998-12-01},
  project = {ant},
  jsubject = {www},
  number = {98-687},
  month = dec,
  note = {Submitted for publication},
  location = {johnh: folder: xxx},
  keywords = {replica selection, http},
  url = {},
  psurl = {},
  pdfurl = {},
  otherurl = {},
  copyrightholder = {authors},
  myorganization = {USC/Information Sciences Institute}
Copyright © by John Heidemann