Our requests, questions and what we first need to understand at a high level before diving deeper. It may be better to split these off into separate threads, but wanted to get some wholistic, high level views first.
We would love to hear from anyone with experience implementing IPFS and libp2p for large date sets and transfers, particularly real time data and particularly video.
Our initial platform is PC/Windows. Believe it or not, thatās for security (PlayReady 3 in hardware and SGX) as well as the massive installed base of PCās. Are there any plans for a Universal Windows Platform (UWP) app environment version of IPFS?
Can we āmergeā IPFS with our Secure Peer Assist (SPA) architecture? One specific and high priority requirement is that our architecture uses a combination of P2P and CDN to āfill gapsā. This allows us to implement a āmost needed packetā rather than ārarest packetā algorithm. This is very useful for real time or near real time playback of video and for seeding. It also combines the strengths of centralised and decentralised storage and provides a role for all in future. We are partnered with EdgeCast (Verizon).
Our video files are packaged with Dynamic Adaptive Streaming over HTTP (DASH) using Common Encryption (CENC). Hopefully that isnāt a problem?
It seems that IPFS encrypts the source files at rest. Is this correct? That is good, but we use encryption that is approved by the studios. Specifically, we need to support PlayReady 3 (PR3) SL3000 encrypted video segments, packaged in DASH CENC and fMP4 containers. We need to understand how IPFS encryption affects and works with PR3.
Very interested to discuss āroutingā options. We have some ideas there.
Can we implement a secure sub-community within the larger IPFS community (weād prefer not to do this but see next question).
How can we ensure no āundesirableā content is stored on our usersā IPFS ābesideā Hollywood content? E.g. pirate movie BitTorrent files or porn. This is a āperceptionā issue, given this happens already on shared environments, but it is a very real and high priority concern of Hollywood, especially some of the security people in the more conservative studios. Weāve seen a bit of discussion around this but havenāt found a clear and definitive answer yet.
We very much like the open transport model. If we provide reliable transport, optimised for large file transfer, presumably that can use UDP?
A little off topic, but highly related (and we canāt find a forum): we love Filecoin. It solves our problem of āincentingā disk sharing and file sharing and creates a currency for our platform. Love their work. Does anyone know if we can create our own, branded version? Even just in name only? The underlying purpose of Blust+SPA is the same, but it would be good to have a currency that reflects our application.
We havenāt been able to find any official Filecoin forums. We would love to participate if anyone can point us to one. Particularly around the economic/business models being contemplated.
Given positive answers to the above (which seems likely), how do we work together to build the worldās first legal, approved, peer to peer movie (and other) file sharing system?
in IPFS you canāt remove a content but you can unpin a ressources if all peer unpin a content the ressources is not up in your swarm. so you can use bot for check content of ressources and unpin a ressource if contents is āundesirableā.
Thanks Josselin. Weāve seen some discussion around the ability to choose what you share. Do you know if this has been implemented anywhere? This is VERY important for the studios.
With IPFS every user sharing only the data that they choose to download beforehand. If your users are the Hollywood studios and those studios donāt download any porn they wonāt store any porn either.
Itās similar to how a company that hosts linux distributions via bittorrent wonāt store any porn on their PC either. At the same time, every user that actually downloads linux distributions from them is free to also download porn via bittorrent and thatās outside of the control of the company that hosts the linux distributions.
Our customers are people who download the Blust app to watch movies. Hollywood supplies us with those movies and they set the terms. So if we control the app and IPFS and the user just chooses how much they share, we can prevent other content being shared on the IPFS associated with the Blust app? It canāt be overridden? That would be perfect. Obviously users can download other IPFS apps but thatās out of our control. Long as we can prevent insertion of undesirable content in the file systems under our control.
iām not sure but you can test locally to create to application which use IPFS in App A define a specific repo_path (ex ~/repo_app_a) and App b another repo_path (ex ~/repo_app_b) in IPFS network it can two speprate node so with two seperate DHT. After that imagine the app A is blust you can in this dht for example exlude all peers else your peers node and hollywood provider. i you add a ressource on app A et after you try to get that with http://ipfs.io/ipfs/<your_hash> (if your peers and hollywood reject all also) that not exist i guess but if you to do that on app B your hash is gettable by ipfs.io.
I donāt know how official it is, but several of Protocol Labās devs are mods at the filecoin subreddit: Reddit - Dive into anything. Itās the most official forum Iāve come across.
I donāt think that is correct. Iām not aware of any encryption for blocks at rest (i.e., in the IPFS repo). If data is encrypted using a proprietary product I donāt see why that would cause problems; IPFS should just store whatever you give it.
Can you access: http://blust.tv ok? The primary domain for the file that raised the Phising alert is windows.net (Azure blob storage default). What Metacert product were you using that tripped the alert?
