BitTorrent: The story of neutral open source technology and piracy

As a representative decentralized file storage protocol in the "pre-blockchain era", how can BitTorrent and its story of development and decline inspire us?

Words: Affe

What is BitTorrent

BitTorrent is a classic peer-to-peer protocol developed by American programmer Bram Cohen.

To put it simply, a complete file in BitTorrent is divided into many file puzzles, which are scattered across the personal computers of individual downloaders. If you want to get a complete file, the BitTorrent client will help you ask who has the file puzzle you need on their computer and download it for you, and if the puzzle you download is needed by someone else, the BitTorrent client will also upload the file puzzle you have for others to download.

Therefore, the more people who download the same file, the faster it will be downloaded (because each downloader can collect several file puzzles at the same time and can make full use of their upload network bandwidth). It's essentially a "all for one, one for all" sharing network. However, in such a network, there is a component similar to a centralized node: the Tracker server. We know that the BitTorrent client can help you get resources from other people's computers, but how does it know "which people's computers have which file puzzles"? That's where the Tracker server comes in. The Tracker server records information about which clients own which file puzzles, and every time a BitTorrent client wants to download a file, it asks the Tracker server where to find the corresponding file puzzle.

There are three points worth noting here: first, the downloader is also responsible for uploading the resource for others to download while downloading the resource, otherwise everyone will only download but not upload, and everyone will have nothing to play. Second, the BitTorrent protocol relies on a relatively centralized Tracker server to function, and if the server doesn't work properly, then all clients that depend on that server won't work either. Third, the main use case of the BT protocol is to share all kinds of audio and video resources, a considerable part of which are pirated resources (Pirating, and "piracy" ). From these three characteristics, we can better understand the challenges encountered in the popularization of the BitTorrent protocol.

"It's piracy": regulation sanctions piracy

After a large number of users used BitTorrent to share pirated movies, the Motion Picture Association of America got into trouble. There is no doubt that BitTorrent facilitates the spread of pirated film and television. Despite Bram's insistence on "providing only the technology and not the content", and according to his wife, "Bram is probably the only one of the BT users who doesn't download any video and pirated software without torrenting, the sheer amount of copyright infringement on BT has attracted regulatory attention, and governments around the world have restricted the use of BT to spread piracy, and the most straightforward way is to shut down the well-known Tracker servers."

In the case of BitTorrent, getting pirated videos and for free is the first thing many people think of using torrent. And the emergence of each P2P technology is more or less a circumvention of "regulation". This "underground" nature is both the charm of BT and the inevitable reason for the regulatory crackdown.

I won't go into too much detail about whether we should push for more open copyright laws, such as the well-known free software movement and the concept of copyleft, which is already a cliché in the blockchain world.

Interestingly, BitTorrent's technical characteristics, which "rely on centralized tracker servers to run", also provide an excellent means of cracking down on regulation. This indirectly led to the adoption of technologies such as DHT that can avoid the dependency of a single server, such as IPFS, which will be discussed later, and the blockchain ecosystem as we know it today.

Download only and don't upload: the bloodsucking of centralized service providers

In addition to regulatory factors, the existence of centralized service providers has also made "seeders" more and more rare. "Seeding" refers to the act of downloading a file and at the same time uploading parts of the file for others to download (all for one, one for all). However, some centralized service providers, such as the famous network disk and the famous thunder, will establish special centralized BT resource stations, but these resource stations do not contribute upload traffic to the entire BT network in the process of downloading, and occupy a large number of upload traffic of individual clients. After collecting a large number of resources contributed by individuals from the BT network, these resources can be made into "membership rights" or "offline download" functions for users to make money. Since it is the individual user who pays the network fee for uploading the broadband, it can be considered that these centralized resource stations "prostitute" the contribution of individual nodes to the network. Due to the asymmetry of download and upload traffic, other torrent client users who comply with the "download also upload" will gradually move to these centralized service providers due to the slow download speed. As a result, the activity of the entire network has deteriorated.

Of course, such a process takes place without the perception of most ordinary users. The average user doesn't know what's going on behind the BT protocol, where the resources are downloaded from, and what role their computer plays; They also have no incentive to pay attention, and most of them aim to download free movies and don't care about the philosophy, technical details, and operating principles behind the technology. Before I studied computer science, I didn't understand why some movies download slowly and others download fast (answer: the more people who download at the same time, the faster the download).

This is a classic "tragedy of the commons": due to the imperfection of the incentive mechanism, malicious competitors can ignore the concept of "all for one, one for all" to prostitute BT resources. Even without looking at the development of BT networks from an ethical perspective, we can see that a well-designed incentive/punishment mechanism is actually an essential component of a P2P network if it is to survive in the long term.

Niche Hobbies: The Current State of Private Tracker Sites

In addition to the above-mentioned regulation and blood-sucking, the availability issues caused by the real-world network topology, the code architecture of the protocol itself, the interactive experience of the protocol, the blocking of BT ports by network operators, and the competition from cheap and high-quality centralized streaming media are all reasons why BT has evolved from a highly-anticipated file interaction protocol to a niche hobby today. There are PT (Private Tracker) sites all over the world, and unlike the original open environment, today's PT sites are extremely strict about new people joining. Take Mantou Station (a well-known PT site) as an example, if you don't have an invitation code, the monthly "donation" (actually subscription) fee is 30 USD, and the lifetime is 520 USD. Some sites are not open for subscriptions at all, and can only be accessed by invitation.

The Chinese PT circle has long had a set of "black words" that all niche hobbies have, and in each PT station, there are systems that need to be strictly followed to avoid doing evil or attracting unnecessary attention, and there are mature mechanisms to encourage "seeding". The once "decentralized" BT network community has also had to establish a partial centralized trust framework to continue to develop "underground". This is undoubtedly an interesting story in today's blockchain world." "Centralization" and "decentralization" are by no means two discrete ideas. There is a very specific spectrum between them: the details of the protocol, the specific implementation, the influence of the external environment, etc., will have an impact on the positioning of a decentralized network on the spectrum.

Interestingly, although the architecture of a single PT site is most likely centralized, each PT site usually has an "official kiln" area, which is an invitation link to other PT sites. These local centers established through strong trust have produced a structurally "decentralized" network in a spontaneous way, which is a particularly memorable event in the author's opinion.

Summary

From a technical point of view, due to the existence of Tracker in BitTorrent, which is easy to be blocked by a single point, subsequent decentralized storage protocols have learned this lesson. In addition, since BitTorrent is a file storage, its main purpose is to store multimedia data as a blob block storage, which is very different from the OLTP used in the application layer and the OLAP system used in the data analysis layer. In a follow-up article, we'll take a look at how IPFS compares to BitTorrent. The author did not delve into the technical details and did a first-hand investigation with the community when researching BitTorrent, so some of the facts and technical details may be incorrect

Reference

why BitTorrent is not popular in the mainland

Summary of well-known PT sites at home and abroad

Introduction to IPFS

PT circle vernacular

BitTorrent Wikipedia

Father of BT: Piracy, not my original intention