\documentclass[11pt]{article}

% preamble

\usepackage{a4wide}
\usepackage{mathptm}
\usepackage{epsfig}
\usepackage{alltt}
\usepackage{helvet}

\makeatletter
\def\verbatim@font{\normalfont\sffamily\fontencoding{T1}\selectfont
    \let\do\do@noligs
    \verbatim@nolig@list}

\renewcommand{\ttdefault}{\sfdefault}
\usepackage{url}
\def\UrlFont{\sf}


\setcounter{secnumdepth}{4}

\newcommand{\DAN}[1]{[{\it {#1}}] \\ }
\newcommand{\status}{{\it Status:}}


\bibliographystyle{acm}


\begin{document}

\begin{tabular}{ll}
Document: & {\bf Tribler Design Note} \\
Subject:  & {\bf YouTube Decentralization } \\
Author: & {\bf Yuan Yuan} \\
Date: & {\bf \today} \\
Status: & {\bf\it DRAFT} \\
\end{tabular}

%
\section{Overview}

Now only a few part of contents which users can download in Tribler
are legal things, most of them are illegal. YouTube is a famous
video website. All the videos in YouTube are legal. In order to
import more legal content into Tribler, we can automatically collect
the video information from YouTube, download the video which user
want to watch into disk  and generate the torrent file for the
video. Thus when other users want to watch the same video while
having the torrent file for that video, they can download the video
from P2P world and don't have to connect to YouTube again. We call
this YouTube Decentralization. This document specifies the overall
process for YouTube Decentralization.

\section{Video Information Collection}

The first step in YouTube Decentralization is to automatically
collect the video information from YouTube and save them into
database. Including video name, url, picture, tags, category and
added time, etc.

\section{Video Play}

When a user want to play a certain movie, the program will connect
the YouTube to download the video file and play it for the user.
When the play is finished, the video file will be saved in the
user's disk.

\section{Torrent Generator}

For each video file saved in the disk, there will be a torrent file
generated. The torrent file contains the information about the
video. Such as video name, url, picture, tags, category and added
time, etc. Two users can exchange torrents using some methods like
BuddyCast(?).

\section{P2P download}

When a user want to play a certain video which is not located in the
user's disk, the program first try to find the torrent file for this
video. If such torrent file can't be found, repeat the step ''Video
Play'' and ''Torrent Generator''. But If the torrent file can be
found, the program will connect the other users who have the video
and download it using P2P technology.

\section{Distributed Tracker}

There will be no official tracker for these torrents. So we will use
"Distributed Tracker" instead normal tracker.

\section{Advantage and Disadvantage}

The advantage of this program is that it can automatically
decentralize the YouTube website and put it into the P2P world.

The disadvantage is that the speed is much lower than YouTube. In
YouTube, you can start to watch the video in few seconds after you
click the video-URL if you have a good internet connection. But in
P2P world, you may wait few minutes before you can start to watch
the video. It is Because the advantage of P2P downloading is based
on the large file downloading - the file is divided into many pieces
, you can download different pieces from different seeders or
leechers. But the videos in YouTube are so small (8.3MB Per Video,
based on the average length of the top 100 downloads of all time).
For these small files, the connection time and bandwidth of content
owner is more important than the number of content owner. But most
of the ''seeders'' in P2P world are personal computer, they can't
offer the large internet bandwidth and computing power as YouTube
server.

The other problem is that YouTube is a business company, is it legal
to these things?

\end{document}