Performance bounds for peer-assisted live streaming

Abstract

Peer-assisted streaming is a promising way for service providers to offer high-quality IPTV to consumers at reasonable cost. In peerassisted streaming, the peers exchange video chunks with one another, and receive additional data from the central server as needed.In this paper, we analyze how to provision resources for the streaming system, in terms of the server capacity, the video quality, and the depth of the distribution trees that deliver the content. We derive the performance bounds for minimum server load, maximum streaming rate, and minimum tree depth under different peer selection constraints. Furthermore, we show that our performance bounds are actually tight, by presenting algorithms for constructing trees that achieve our bounds.

Point-to-Point and Congestion Bandwidth Estimation: Experimental Evaluation on PlanetLab Data

“…This assumption is realistic in many scenarios concerning large scale Internet platforms (like for example when the system consists of home computers connected to the network by DSL connections), hence this model has been used in several studies to design or analyze communication algorithms for video-on-demand [10], peer-assisted streaming [2] or master-slave tasking [11]. The LastMile model assigns to each node i an outgoing bandwidth b out i and an incoming bandwidth b in i , and then the available bandwidth between two nodes BW (i, j) can be computed easily:…”

“…In the case of delivering video on demand [1], or performing peer-assisted streaming [2] for example, estimations of available bandwidth allow the construction of an efficient overlay topology. Efficient algorithms for broadcasting or for organizing master/slave communications have been proposed [3], [4], and they rely on an accurate representation of the performance of the network.…”

“…Then, in Section IV, we go beyond estimation of point-topoint performance and address the problem of congestion: is it possible to predict the performance obtained when several communications take place at the same time? The Last Mile model, used as a communication model in several algorithmic studies [10], [2], [11], actually specifies that communications can happen in parallel, as long as the outgoing or incoming bandwidth limits of each node are satisfied. In Section IV-A, we present our dataset, which contains dedicated measurements performed on PlanetLab specifically to study this question.…”

Broadcasting on Large Scale Heterogeneous Platforms under the Bounded Multi-Port Model

“…However, none of them has considered the constraint added by the presence of firewalls in the system. The work closest to our approach is by Liu et al [24] in which they provide bounds for the streaming rate, the upload rate of the source needed to ensure a given stream rate, and the depth of the distribution trees produced. Degree constraints are also considered in their work, but with specific limitations.…”

Resilient Tree-Based Video Streaming with a Guaranteed Latency

“…According to the popularization of video sharing services such as YouTube and Hulu and the performance limitation of content delivery networks (CDNs) due to the intrinsic bottleneck at the edge servers, Peer-to-Peer (P2P) video streaming has attracted considerable attention in recent years. 1,3,4,[7][8][9][10]13,14,[16][17][18][19][20][21]26,30,31 The report on the demonstration experiments conducted by NHK science & technology research S. Fujita laboratories during London Olympics in 2012 shows that P2P video streaming realizes the delivery of live contents (e.g., the final of the men's singles of tennis tournament) to more than 1600 subscribers in 1.5 Mbps in a stable manner. 22 In this paper, we focus on the problem of preserving the performance of such a P2P video streaming sufficiently high, with respect to the latency, stability, frequency of playback suspension, and the waiting time due to playback suspension.…”