Font Size: a A A

Design And Implementation Of Network Coding Based P2P Streaming VoD Systems

Posted on:2011-02-21Degree:MasterType:Thesis
Country:ChinaCandidate:L X GaoFull Text:PDF
GTID:2178360305997301Subject:Computer applications and technology
Abstract/Summary:PDF Full Text Request
The application of live multimedia streaming VoD (Video-on-Demand) systems is getting more and more focus with the rapid growth of Internet these years; and the P2P (Peer-to-Peer) mode has already become the most popular pattern instead of the original one based on C/S (Client & Server). With P2P, systems can easily meet the requirement of mutual cooperation using every idle bit of peers'upload bandwidth. So the scalability of the systems is greatly increased, which in turn is the biggest disadvantage of C/S mode.However, the churn of network situation, together with the shortcoming of transmitting pattern, will slow down the P2P streaming systems. Recently, the concept of Network Coding has been introduced into such systems, which was proposed for the first time in 2000 by R. Ahlswede et al. Network Coding allows the intermediate nodes in the network to take operations on some incoming packets to combine them together, instead of originally copy and forward, which makes the whole system more effective. In P2P systems with Network Coding, all the encoded blocks are treated equally, which makes it easier to implement push-based cooperative transmission and achieve higher transmitting rate. Meanwhile, Network Coding greatly reduces the complexity of cooperation scheduling, as well as provides robustness when network status is bad or fluctuating. The process of Network Coding will bring issues about computing performance. Luckily we can use NVIDIA's CUDA platform to accelerate the encoding, as well as to enlarge system throughput.This article studies the problems emerge when Network Coding is introducing into P2P-based streaming VoD systems, and then designs and implements a P2P-based streaming VoD system with Network Coding, SONICVoD. We also implement the GPU accelerating module on the server end. It discussed all the challenges we meet in the whole procedure, and provides decent solutions. After the large-scaled performance evaluation, we gain a lot of useful data, with which we can make deeper researches. We also verify the important conclusions from previous work about P2P streaming systems with Network Coding, as well as give solid proof of our SONICVoD's effectiveness and efficiency.
Keywords/Search Tags:network coding, P2P, streaming systems, Video-on-Demand, GPGPU
PDF Full Text Request
Related items