A simple tree management algorithm is presented that provides the necessary path diversity and an adaptation framework for MDC based on scalable receiver feedback is described, which shows very significant benefits in using multiple distribution trees and MDC, with a 22 dB improvement in PSNR in some cases.
Abstract:
We consider the problem of distributing "live" streaming media content to a potentially large and highly dynamic population of hosts. Peer-to-peer content distribution is attractive in this setting because the bandwidth available to serve content scales with demand. A key challenge, however, is making content distribution robust to peer transience. Our approach to providing robustness is to introduce redundance; both in network paths and in data. We use multiple, diverse distribution trees to provide redundancy in network paths and multiple description coding (MDC) to provide redundancy in data. We present a simple tree management algorithm that provides the necessary path diversity and describe an adaptation framework for MDC based on scalable receiver feedback. We evaluate these using MDC applied to real video data coupled with real usage traces from a major news site that experienced a large flash crowd for live streaming content. Our results show very significant benefits in using multiple distribution trees and MDC, with a 22 dB improvement in PSNR in some cases.
TL;DR: This paper presents Bullet, a scalable and distributed algorithm that enables nodes spread across the Internet to self-organize into a high bandwidth overlay mesh, and finds that, relative to tree-based solutions, Bullet reduces the need to perform expensive bandwidth probing.
TL;DR: In this article, the authors describe principles in designing multiple description coding (MDC) video coders employing temporal prediction and present several predictor structures that differ in their tradeoffs between mismatch-induced distortion and coding efficiency.
TL;DR: It is shown that the main factors attributing in the inferior performance of the tree-based approach are the static mapping of content to a particular tree, and the placement of each peer as an internal node in one tree and as a leaf in all other trees.
TL;DR: The main design goal of PRIME is to minimize two performance bottlenecks, namely bandwidth bottleneck and content bottleneck, and it is shown that the global pattern of delivery for each segment of live content should consist of a diffusion phase which is followed by a swarming phase.
TL;DR: The basic taxonomy of peer-to-peer broadcast is described and the major issues associated with the design of broadcast overlays are summarized, including the key challenges and open problems and possible avenues for future directions.
TL;DR: Results from theoretical analysis, simulations, and experiments show that Chord is scalable, with communication cost and the state maintained by each node scaling logarithmically with the number of Chord nodes.
TL;DR: The concept of a Content-Addressable Network (CAN) as a distributed infrastructure that provides hash table-like functionality on Internet-like scales is introduced and its scalability, robustness and low-latency properties are demonstrated through simulation.
TL;DR: The potential benefits of transferring multicast functionality from end systems to routers significantly outweigh the performance penalty incurred and the results indicate that the performance penalties are low both from the application and the network perspectives.
TL;DR: This measurement study seeks to precisely characterize the population of end-user hosts that participate in Napster and Gnutella, and shows that there is significant heterogeneity and lack of cooperation across peers participating in these systems.
TL;DR: It is found that forwarding packets via at most one intermediate RON node is sufficient to overcome faults and improve performance in most cases, demonstrating the benefits of moving some of the control over routing into the hands of end-systems.
Q1. What contributions have the authors mentioned in the paper "Resilient peer-to-peer streaming*" ?
The authors consider the problem of distributing “ live ” streaming media content to a potentially large and highly dynamic population of hosts. Their approach to providing robustness is to introduce redundancy, both in network paths and in data. The authors use multiple, diverse distribution trees to provide redundancy in network paths and multiple description coding ( MDC ) to provide redundancy in data. The authors present a simple tree management algorithm that provides the necessary path diversity and describe an adaptation framework for MDC based on scalable receiver feedback.
Q2. What is the main argument for using source coding?
Several researchers have advocated the use of source coding, possibly in conjunction with path diversity, to make data transfer robust to packet loss.
Q3. When did the use of multiple description coding in conjunction with multipath routing in (telephone?
The use of multiple description coding in conjunction with multipath routing in (telephone) networks dates back to the late 1970s [19].
Q4. How do the authors insert a new node into a fertile tree?
To insert the new node into its fertile tree, the authors start at the root and proceed down until the authors reach a level that either has a node with room (i.e., with spare bandwidth) or a node with a sterile child.
Q5. What is the expected distortion of the mth packet?
Given p(m) and the operational rate-distortion function D(R), this expected distortion can be minimized using a simple procedure that adjusts the rate points R1, . . . , RM subject to a constraint on the packet length [17], [33], [26]
Q6. How long does it take for a tree to be repaired?
E. Impact of Repair IntervalThus far the authors have assumed that it takes 1 second for a tree to be repaired following the departure of a node.
Q7. What is the way to scale up a tree?
Should the tree management processing on one root node become a bottleneck, it would be easy to scale upusing a (possibly distributed) cluster of roots and directing each client to one of the roots, say at random.
Q8. What is the average quality of the received stream at the clients?
The authors report the average quality of the received stream at the clients, quantified using the Peak Signal-to-Noise Ratio (PSNR) metric, which is computed from the luminance distortion D: PSNR = 10Log10(2552/D).
Q9. How does the PSNR curve change as the client population drops?
Soon after that, PSNR spikes up as the client population drops, to the point where almost all nodes can directly become children of the root and hence experience little disruption.
Q10. How can the root save the network round trip?
(The network round trip needed for this check can possibly be saved by having the parent piggyback its packet loss rate information on the data stream it forwards to its children.)
Q11. What is the p(m) distribution for each GOF interval?
For each GOF interval, the tree manager computes the p(m) distribution (Section III-A) corresponding to the number of descriptions received over all clients.