Resilient tree-based live streaming in reality

Research output: Contribution to book/Conference proceedings/Anthology/ReportConference contributionContributedpeer-review

Contributors

  • Benjamin Schiller - , Technische Universität Darmstadt (Author)
  • Giang Nguyen - , Technische Universität Darmstadt (Author)
  • Thorsten Strufe - , Technische Universität Darmstadt (Author)

Abstract

Peer-to-Peer (P2P) streaming has been a scalable and cost-efficient solution to deliver live video streams. To reduce the bandwidth demands at video servers, it uses resources (mostly upload bandwidths) of participating peers. However, depending on unreliable end-hosts to build an efficient and robust streaming system is a difficult task. Various approaches have been proposed. They are often classified by the type of overlay topologies and the way video packets are disseminated in the overlay. Single-tree-push systems have low latency, but are less resilient to network dynamics. Mesh-pull systems are more resilient to failure and churn, but suffer larger delays and overhead. Hybrid systems are based on a mesh-pull system. Peers in these systems may switch between pull and push modes to deliver video packets at certain conditions. However, their resistance against Denial-of-Service (DoS) attacks is unknown. Multi-tree-push systems split the video stream into stripes, and deliver them using disjoint trees. The latency is therefore low, which is suitable for a live streaming service. Moreover, it is also proven resilient against DoS attacks via formal analysis and simulations [1]. Nonetheless, it is not known how the system performs in real world conditions.

Details

Original languageEnglish
Title of host publication13th IEEE International Conference on Peer-to-Peer Computing, IEEE P2P 2013 - Proceedings
PublisherIEEE Computer Society
ISBN (print)978-1-4799-0521-8
Publication statusPublished - 2013
Peer-reviewedYes
Externally publishedYes

Publication series

SeriesInternational Conference on Peer-to-Peer Computing
ISSN2161-3559

Conference

Title13th IEEE International Conference on Peer-to-Peer Computing, IEEE P2P 2013
Duration9 - 11 September 2013
CityTrento
CountryItaly

External IDs

ORCID /0000-0001-7008-1537/work/160479874

Keywords

ASJC Scopus subject areas