A Context-Aware HTML/XML Document Transmission Process for Mobile Wireless Clients

  • Authors:
  • Huamin Chen;Prasant Mohapatra

  • Affiliations:
  • Department of Computer Science, Engineering II, University of California, Davis, USA 95616;Department of Computer Science, Engineering II, University of California, Davis, USA 95616

  • Venue:
  • World Wide Web
  • Year:
  • 2005

Quantified Score

Hi-index 0.00

Visualization

Abstract

Delivery and rendering of HTML/XML documents has been a core task in many contemporary networking applications. In mobile wireless networks, efficient handling of these types of documents is necessary due to the frequent disconnections, packet loss, and high bit error rate. One approach to address the challenge is the ability to process and reuse the partial data. Current application protocols like HTTP cannot support this approach due to the following constraints: TCP's in-order data uploading to the applications and tag matching. We propose a context-aware transmission process (CATP) to run on top of UDP. This protocol does not transmit HTML/XML files in-order. Instead, it reorganizes the files and transmit tags first before transporting their enclosed data. Conforming browsers receive the file structures and fill in with subsequent data packets in whatever sequence they arrive. As a result, lost and delayed packets do not hinder rendering of those that are logically behind but have already arrived at the client sides. Thus the retransmission of the lost frames can be concealed and overall user perceived performance improved. The user-perceivable performance is quantified in terms of silent time during which no activity is observed at the browser display. The protocol also facilitates partial content caching, amortizing network transmission overhead, and non-interactive applications of Web services. We validated this protocol through prototype implementation and compared the performance with TCP and in-order delivery UDP schemes. Our protocol provides better user-perceivable performance under various loss rates and document sizes.