Load management in distributed video servers
ICDCS '97 Proceedings of the 17th International Conference on Distributed Computing Systems (ICDCS '97)
Can internet video-on-demand be profitable?
Proceedings of the 2007 conference on Applications, technologies, architectures, and protocols for computer communications
ECHOS: edge capacity hosting overlays of nano data centers
ACM SIGCOMM Computer Communication Review
Analysis of bittorrent-like protocols for on-demand stored media streaming
SIGMETRICS '08 Proceedings of the 2008 ACM SIGMETRICS international conference on Measurement and modeling of computer systems
Challenges, design and analysis of a large-scale p2p-vod system
Proceedings of the ACM SIGCOMM 2008 conference on Data communication
A framework for lazy replication in P2P VoD
Proceedings of the 18th International Workshop on Network and Operating Systems Support for Digital Audio and Video
Optimal content placement for a large-scale VoD system
Proceedings of the 6th International COnference
Balancing throughput, robustness, and in-order delivery in P2P VoD
Proceedings of the 6th International COnference
Push-to-Peer Video-on-Demand System: Design and Evaluation
IEEE Journal on Selected Areas in Communications
Popularity decays in peer-to-peer VoD systems: Impact, model, and design implications
Computer Networks: The International Journal of Computer and Telecommunications Networking
Hi-index | 0.00 |
In recent years, Peer-to-Peer assisted Video-on-Demand (P2P VoD) has become an effective and efficient approach to distribute high-quality videos to large number of peers. In a P2P VoD system, each peer contributes storage to store several videos to help offload the server. The replication strategy, which determines the videos to be stored at each peer's local storage, plays an important role in system performance. There are two approaches: (a) solve a huge combinatorial optimization problem and (b) use simple cache replacement algorithms, such as Least-Frequently-Requested (LFR) or FIFO. The first approach needs to collect a large number of parameters whose values may be changing, and use some approximation method (such as linearization) to solve the optimization problem, both aspects have accuracy issues. In the second approach, a peer replaces some video in the cache with the currently viewed video, based on local information. While it is simple, we show their performance can be improved by a little centrally collected state information. Specifically, the needed feedback information is the current downloading rate provided by peers for each video. In this paper, we describe a hybrid replication strategy, and give detailed description of how the server collects and maintains the feedback information, and how peers use that information to determine what videos to store and indirectly control their uplink bandwidth contribution. This explains why the hybrid strategy is much simpler and more practical than the combinatory optimization approach. We then use simulation to demonstrate how our scheme out-performs the simple adaptive algorithms. Our simulation results also demonstrate how our scheme is able to quickly respond to peer churn and video popularity churn.