Brightnet: OFF System

Notes from the I2P forums

Let’s say the riaa (or mpaa) wants to bust you. They starts a new node, find the url of a copyrighted content, and download it. They take note of all the blocks they receive, needed to build back the file. Then start monitoring your request. If you request one of those blocks, you have 50% probability that you are downloading the copyrighted content. (the other 50% is if you’re downloading something else that uses that block for xor). If you request two blocks you’ve 66%, and so on. But even in the point of downloading a copyrighted OFF-Url, the node of the RIAA could accouse all the IP adresses, it gets ANY block from.

The Rebuild of the File is done locally on your machine. Then they’d have to monitor YOUR side of the network, which means they will have to either break in to your computer or sniff your connection somewhere between your house and your ISP. Even if they do that, one block is not a 50% probability, one block can belong to 100s of files, there is no way to tell. Even if you get ALL of the blocks needed for a file it doesn’t mean that you actually retrieved (press “Save” to your hard disc) that file. Maybe you retrieved another file that was “target-stored” with that file (one file in another), or all the file’s blocks just happened to be in the inserters block-cache when (s)he inserted another file and all of that file’s blocks were used as randomizers. So you can insert a copyright-free Linux Distribution into a block of a copyrighted movie. But if you then download the Linux-Distribution with the OFF-Link and need one mixed-in block of the copyrighted movie - what is the problem to download a Linux Distro with a randomized and XOR-ed Block of another file, no one can refer to (because it is XOR-ed)? For the tracking of the Postmans it is nearly the same: the Riaa can download the copyrighted Media and accuse all IP adresses, it gets blocks from. But what would then be the accusation? To transfer the media to the riaa-node, which is random data? You see, the MEANING is the essential thing for random data, and this is the OFF-Link. The riaa can only accuse the ones, which gave the RIAA the OFF-Link. And if the RIAA itself is uploading a copyrighted media as a honeytrap, they need to download the OFF-Link itself to them. But of course the conclusion is then, that no one should download the OFF-Links of the RIAA. You are safe in downloading here again, but overgiving the OFF-Link might be risky, but does the Riaa know, what you do with the link? as there are good chances, that you do nothing with it. There is no relationship between all… in any case.

The URLs are the weak point of the system, but this is designed to be fixed with turtle hopping.

Unfortunately, even if you choose to relay, but hide your OFF-Url-Link-URLs, it is conceivable that some wily lawyer would misinterpret this as causation and single you out for attack. If they seized your machine then your collection of URLs would be exposed as potential evidence to a jury that didn’t understand any of this.For that reason, that searching OFF-Url-Links in the OFF-System is not anonymous, we discuss the development of a search system sperated from the OFFSYSTEM (e.g. by turtle hopping searches in a friend-to-friend network).

Anyway, don't trust OFF (nor any of its descendants Monolith, OFFLOAD and Blocksnet) YET for anonymity of data. Experimentation is fine tho.