Introduction/Abstract

iTorrent is essentially a fusion of iTunes and BitTorrent. It works by using a centralized BitTorrent client with the basic infrastructure of iTunes and the iTunes Store, which takes more of the burden of uploading files to customers off of the iTunes central server. This is beneficial in that it makes the process of uploading and downloading media faster and more efficient. Although customers will be charged to use iTorrent, there is an incentive system which makes it more user friendly. This incentive system will either provide monetary discounts for users who allow their files to be uploaded by other users or will involve a point system, which will provide users with iTunes credit in accordance with the amount of megabytes that they upload to fellow users. This will be something like $1 toward the iTunes store for every 10,000 megabytes uploaded. Although services like Napster have been held accountable for their clients' actions due to the RIAA and MPAA demanding that they share information about their clients with them, iTunes and therefore iTorrent had less of a claim to neutrality and would most likely not be handicapped by legal proceedings against it. Technically speaking, there are two possible models for combining BitTorrent capabilities with iTunes. One of these is to distribute movie files through current BitTorrent clients. This would work by way of people downloading files from their peers and subsequently acquiring a license from the iTunes Store to be allowed to play those files. The second way that iTorrent could work is by Apple incorporating a BitTorrent client framework of iTunes. It would be activated as iTunes opens and be used for iTunes-provided content only. iTunes would use a BitTorrent 'tracker' that would connect them to other users and begin to transfer files through an initiation with the iTunes store. There are a few areas of concern with iTorrent. One involves the robustness of the program. Apple would need to control the system and constantly maintain its quality, which could be difficult, especially with the first technical model. Any lack of quality would have to be made up by iTunes. Another possible road bump would be the necessity of a certain density of files on the network for necessary speed and connection stability. The central server would have to bear the burden for any lack of file density at first. Ethically speaking, the only possible area for concern with iTorrent would be the incentive system. Those same individuals in the entertainment industry who believe that all media should be purchased may be concerned with the fact that after accumulating a certain amount of discounts or points, users essentially are able to download selevted media for free. However, iTorrent would still maximize its profits through heavy use of the incentives that are provided.

Business Overview

Recently, the increased demand for downloadable movies and television shows has greatly increased the strain on iTunes’ central servers. As more and more users demand these large files, it seems likely that iTunes will have no other option than to increase the space available on its central servers. However, by including a BitTorrent-like client into its existing iTunes client, iTunes can spread the burden of uploading these large files to its users, thereby forgoing the cost of additional server space.

The system we plan to propose to Apple incorporates a centralized BitTorrent client into the existing framework of the iTunes store with the purpose of relieving iTunes’ central server of much of the burden of uploading large files directly to purchasers. When a movie is purchased, the iTunes client will (if it determines an appropriate number of users are available to upload the file) connect the purchaser to these nodes through which the movie will be downloaded.

In order to make this system attractive to users, it is necessary to produce some sort of incentive system. The two ways we envisioned the system working are (a) offer the purchaser a monetary discount up front if they agree to allow others to upload the file from them and (b) create a system by which users earn points for the amount of megabytes uploaded to other users, which can then be converted into credits toward future purchases made at the iTunes store. Because a monetary discount up front would do nothing to prevent users from buying media at the discounted price and then moving the file to another location out of iTunes’ reach, we decided to pursue the second option. In order to make a points system work, iTunes would need to keep track of the amount of bytes uploaded by each user. There are a few issues we need to sort out with the points system: (a) the value of points in dollar/credit terms, (b) limits (if any) we should place on the amount of megabytes a user can upload or limits on the number of points available to earn in a specified period of time, and (c) how we equate the profits lost as a result of point payments (as opposed to dollar payments) and the money saved by reducing the strain on iTunes’ central servers. I will also determine what can be done in the way of marketing to make the possibility of uploading through the iTorrent client attractive to users.

Legal Overview

There are obviously many areas of law that could come into play here. So far, I’ve been searching for a few recent court cases in each area that may help me dive into deeper issues, many of which are very recent and some of which are still undecided. Since so many areas are sort of…wishy-washy, I am guessing that much of the legal brief will be in the form of safety recommendations/advisory warnings against stepping around possible pitfalls.

Areas of Concern:

1) Having a more centralized server:

A lot of lawsuits that have developed against BitTorrent have been against people who host the tracker servers. If iTunes implements a centralized server and tracker, will it be held more responsible for its clients’ actions? One of the major precedents here would be in the lawsuits against Napster. The recording industry argued that the “indexing server” Napster maintained made the company responsible for acts of infringement committed by users. In the past P2P networks have been saved from this due to their decentralized nature.

A centralized server that stores information could also lead groups like the RIAA and MPAA to demand that iTunes share the information with them about customers who are potentially infringing copyright use. In RIAA vs. Verizon, the court held that the DMCA did not authorize the issuance of a subpoena to an ISP acting only as a conduit…but iTunes is not an ISP, and has less of a claim to neutrality with a centralized server. Could iTunes be forced to share valuable client information?

2) The DMCA

From class we’ve seen how poorly this act has been used. I think the main issue here would be whether or not iTunes would be held responsible for any possible “circumventing” that their clients might do, since the use of BitTorrent itself by Apple would obviously be through legal/contracted/purchased terms. I have not found any cases that would indicate that iTunes would have any responsibility, especially since there are things like QTFairUse that strip the DRM from iTunes songs for Linux. QTFairUse has survived and is available for download today, and I could not find any legal actions taken by either Apple or the music industry for rights violations.

3) Others: DRM lawsuits, Family Entertainment and Copyright Act, BitTorrent’s New Release

There are numerous other issues that I am still looking in to, especially since a lot of the legal issues have so far been against music sharing, not video sharing. For example, Window’s Vista is apparently crippled by the DRM restrictions it must follow. Would iTunes videos be subject to any similar media restrictions? Also in a recent case, Sony was sued and found liable for not telling consumers CDs sold by the company contained DRM software that monitored user listening habits. With all the mods of tracking and recording information that our project proposes, iTunes may have to be very careful and clear about explaining what they are tracking and why, perhaps in the licensing agreement. The Family Entertainment and Copyright Act was just a newer law passed about video recording that I thought may be useful to look at. BitTorrent has recently announced that it will be selling movies similar to this project, so it will be interesting to watch that develop over the next few days.

Technical Overview

There are two basic models for applying BitTorrent capabilities to the iTunes Store. First, iTunes could rely on the most basic extension of BitTorrent – simply distributing movie files through already existing BitTorrent clients. This model would consist of users downloading the files from their peers and then acquiring a license from the iTunes Store in order to play the protected files. Though this would be a relatively simple implementation, the lack of control of this system leaves it lacking in robustness and reliability. In short, for Apple to use BitTorrent in an economically viable way and provide the reliability and quality of service that customers demand from a pay-service, Apple must be able to control the system in order to maintain its quality.

Instead, we believe it is necessary for Apple to take a more centralized approach which will greatly increase the robustness and quality of the BitTorrent service provided. Using a partially centralized model, Apple will incorporate a BitTorrent client into iTunes, which will be activated whenever iTunes is running and will be used solely for trading iTunes-provided content. When a user wishes to download a movie, they will initiate the download at the iTunes Store. The store will activate the equivalent of a BitTorrent ‘tracker’ which will connect the user to other users which have the same content and start the file transfers.

A central iTunes server will manage who has purchased licenses to view the movies, and with this list will be able to set up trackers for the iTunes BitTorrent client (“iTorrent”) to use. The advantage of such a system is that iTunes must already store this information, and the iTorrent client can take advantage of the records held on the central server. Furthermore, the client application can be easily controlled to only share properly licensed files with clients who have purchased licenses for the same files through iTunes already. By using this centralized iTorrent client system, user client nodes within the network can easily perform ‘handshakes’ to determine the legitimacy of file transfer requests. Furthermore, this system will be transparent to the user; that is, the user will initiate the download from the iTunes site and will not know whether their download source is the iTunes central server or peer clients.

One main issue that this system will have to address is that of populating the network with these files in the first place. For any p2p client to function, there is the need for seed nodes, which contain the files to be transferred in their entirety. Furthermore, a certain density of files on the network, which I will refer to as a critical mass, is required to ensure download speed and connection stability. In order for the system to perform at high enough standards of quality to ensure customer satisfaction, the iTorrent system will have to be supplemented through the iTunes central server. Basically, this entails monitoring the network to track the prevalence of files on the network. If the file availability is below a certain threshold, the central server will have to bear the burden until the critical mass is restored. Of course, the central server will have to provide much of the content until a critical mass is reached, and even afterwards may have to contribute on and off, as nodes enter and leave the network, and users potentially delete files.

The iTorrent model will also employ a reporting system. The main tasks of the reporting capability are (a) to update the central server as to file availability and (b) to track upload amounts. By tracking the amount of data uploaded, clients will report how much users “shared,” and allow Apple to provide monetary incentives in proportion to the contribution by individual users to the network functionality. This will be something like $1 toward the iTunes store for every 10,000 megabytes uploaded. Actual reward to be determined. This model will include options to allow users to only upload while downloading or while the computer is idle.

I intend to research

1. How to determine the "critical mass" required
2. Verification models
3. Expected user upload quantities to help gauge reward amounts?
4. DRM models for the distribution of protected files from a variety of sources
5. Sharing models to determine what percentage of customers must be sharing to create a robust network.

“And that’s all I have to say about that.”

1: Put your blog in a box

The blog for Bit Torrent Project Group 1 of CS182s Spring '07 is now up and running. But you already know that, because you're reading it. Congrats.

This group includes:

1. Legal: Jessica Wade
   
2. Business: Charlotte Nicholson
   
3. Guru: Eric Pritchard
   
4. Technical: Chris Mavricos

That's really all we have right now.

Stay classy, San Diego.