Cache capacity allocation to overlay swarms

  • Authors:
  • Ioanna Papafili;George D. Stamoulis;Frank Lehrieder;Benjamin Kleine;Simon Oechsner

  • Affiliations:
  • Athens University of Economics and Business, Athens, Greece;Athens University of Economics and Business, Athens, Greece;Julius-Maximilian University of Wuerzburg, Wuerzburg, Germany;Julius-Maximilian University of Wuerzburg, Wuerzburg, Germany;Julius-Maximilian University of Wuerzburg, Wuerzburg, Germany

  • Venue:
  • IWSOS'11 Proceedings of the 5th international conference on Self-organizing systems
  • Year:
  • 2011

Quantified Score

Hi-index 0.00

Visualization

Abstract

Peer-to-peer applications generate huge volumes of Internet traffic, thus leading to higher congestion, as well as higher costs for the ISPs particularly due to inter-domain traffic. The traditional traffic management approaches employed by ISPs (e.g. traffic shaping or even throttling) often lead to a deterioration of users' Quality-of-Experience. Previous works have verified that the insertion of ISP-owned Peers (IoPs) can deal effectively with this situation. The mechanism offers caching while exploiting the self-organizing structure of overlays. Thus, it leads to both improved performance for peers and reduced inter-domain traffic costs for ISPs. In this paper, we study how the available IoP bandwidth capacity should be allocated among the various swarms that it can possibly join. In particular, we identify a variety of factors that influence the effectiveness of Swarm Selection and Bandwidth Allocation, and we investigate their impact on the practically relevant example of BitTorrent, primarily by means of simulations.