1.1 Overview
Since the current Internet was early started in 1970s, many new services and applications have been invented based on its architecture, such as WWW, E-mail, BBS and File Transfer Protocol (FTP). These services all based on the so called Client-Server model, on which the server is placed somewhere on Internet and waiting for client’s request. Upon receiving client’s request, the server starts to provide its service to client.
Client-Server model is intuitive and easy for implementation. However, the server always becomes the bottleneck when the system population grows up on Client-Server model because of the limitation in centralized resources. It is always difficult for a server to provide service to a large number of clients at the same time.
In the past decade, the so-called Peer-to-Peer (P2P) network architecture emerged. In this architecture, every peer acts as both client and server. A peer provides and receives service to/from other peers. That is the main difference between Client-Server architecture and P2P architecture. There are several advantages on P2P architecture, it resolves the server bottleneck problem even if a peer is broken. On the other hand, a peer can find out more resources in peers which are located on different places compared to centralized server(s).
The early application on P2P architecture is file-sharing. Every peer can share its owned file among peers. Without centralized administration, peers share their own files more freely. As a result, P2P becomes the best platform on file-sharing. To improve the file sharing performance on P2P architecture, Bram Cohen invents BitTorrent (BT) [1] which is a new architecture based on P2P file-sharing. The core idea of BT is dividing a file into
many fixed-size chunks so that a peer can download a file from different peers for different chunks simultaneously. This can improve download speed to a very fast way. The concept of “chunks” causes great effect not only in file-sharing but also in many aspects.
In addition to file-sharing, there were many applications devised based on P2P architecture, such as Internet Telephony or VoIP. Skype[5] is probably the most famous VoIP service. It is based on P2P architecture and works under proprietary protocol. It was a very successful VoIP application on nowadays Internet.
Another popular application based on P2P network architecture is Internet Protocol Television (IPTV). The main reason that the IPTV runs on P2P architecture is scalability and provisioning cost. Usually, a centralized server can not serve too many clients at the same time especially for multi-media application such as video streaming. However, IPTV based on P2P can utilize every peer’s upload bandwidth to provide contents to other peers.
The working load in centralized server can thus be reduced and system scalability can be greatly improved.
The “chunks” concept in BT affects P2P IPTV system. The source of video content is always divided into chunks and delivered to a subset of peers, which receives chunks from source and shares it with other peers. There are many P2P overlays proposed to achieve the goal. The tree overlay is the first proposed architecture. Every peer is a node in the tree overlay and it derived video stream from parent and delivered contents to its children. The other one is mesh overlay, in which every peer receives and delivers chunks from/to uncertain peers. The mesh overlay is much easier to implement and robust compared to the tree overlay. However, its start-up delay is much longer than that for tree overlay.
There are quite a few commercial P2P IPTV systems developed in recent years, such as PPStream [3], PPLive [2], SopCast [6] and TVAnt [7]. These systems attract many users.
However, most of them are proprietary systems. They do not release their protocol and scheme on their systems. Some researchers are still curious about the inside of these
systems. Works have been done to analyze the system by tracing their packets [11] and we can have a closer view into these systems.
1.2 Motivation and Purpose
People watching traditional TV only has to push on the remote control and wait for less than 1 second. However, the time duration from selecting an IPTV program to the display of the first video frame, a person usually needs to wait for a while. The duration ranges from few seconds to a couple of minutes [12] based on the network state and different IPTV systems. In addition, on the traditional TV system, a user can browse different channels in a few seconds and select a channel to watch. This feature is not available in IPTV system, especially in P2P IPTV system.
Nowadays, there are many P2P IPTV systems based on mesh overlay available.
Unfortunately, mesh overlay may cause longer start-up delay than other types of P2P overlay. Compared to traditional TV system, there is no special incentive except cost concern for users to watch IPTV.
To make P2P IPTV more mature and overcome the problems described above, in this thesis, we propose a new approach to reduce the start-up delay on mesh based P2P IPTV system. We focus on mesh based system due to its scalability and ease in implementation, so more and more P2P IPTV systems are based on it. We intend to address long start-up delay problem in a real system not just for academia research only. Our goal is to shorten start-up delay in P2P IPTV system so that user can experience the little difference compared with traditional TV system.
1.3 Organization of the Thesis
The rest of this thesis is organized as follows. We introduce P2P IPTV systems in
more detail and long start-up delay problem in Chapter 2. In Chapter 3, we describe our proposed scheme in detail from different point of views. We evaluate our scheme through simulation in Chapter 4. Conclusions are stated in Chapter 5.