Design and analysis of a stable set-up protocol for transcoding multicast trees in active networks

  • Authors:
  • Bart Duysburgh;Thijs Lambrecht;Filip De Turck;Bart Dhoedt;Piet Demeester

  • Affiliations:
  • Department of Information Technology (INTEC), Ghent University, St. Pietersnieuwstraat 41, B-9000 Ghent, Belgium;Department of Information Technology (INTEC), Ghent University, St. Pietersnieuwstraat 41, B-9000 Ghent, Belgium;Department of Information Technology (INTEC), Ghent University, St. Pietersnieuwstraat 41, B-9000 Ghent, Belgium;Department of Information Technology (INTEC), Ghent University, St. Pietersnieuwstraat 41, B-9000 Ghent, Belgium;Department of Information Technology (INTEC), Ghent University, St. Pietersnieuwstraat 41, B-9000 Ghent, Belgium

  • Venue:
  • Journal of Network and Computer Applications
  • Year:
  • 2007

Quantified Score

Hi-index 0.00

Visualization

Abstract

Based on active networking, an advanced streaming service was designed to offer different formats of the same stream using a single multicast tree. To that end, the initial format of the stream that is sent into the tree is transcoded to the other requested formats in the nodes of the tree, based on application level functionality residing in these network nodes. To set up such a multicast tree (including the necessary forwarding state in the nodes) and to install the necessary functionality in the nodes, a number of tree set-up procedures were designed. In this paper, performance aspects of these procedures are investigated: the stability and consistency of the forwarding state during the set-up procedures and the influence of dynamic user behaviour on the multicast tree. This performance assessment is based on a thorough analysis of the different set-up procedures and on simulations of the procedures. Based on the analysis, it is seen that great care must be taken during the set-up procedures in order to avoid interference with the existing streams. Therefore, extensions for the procedures are proposed. Furthermore, the influence of dynamic user behaviour on session state and performance attributes is analysed.