What features would you want in a Torrent 2.0 protocol?
>Batch torrents that automatically add new releases.
>Automatic file patching or replacements for v2s.
>Individual torrent seeder/leecher info.
>>54951925
something like bittorrent sync but
>you can choose what to download
>you can choose what to upload, you can upload files without having to download everything to the drive
>>54951925
Some sort of privacy would be nice. I don't like the idea of everybody in the world being able to see my IP and what it's downloading in the public swarm.
Apart from that, the big problem to solve would be trackerless search and some sort of distributed quality assurance to make sure it isn't 99% mislabelled garbage and botnets.
>>54951925
not sure how one can implement this, but even more decentralization
>depends on trackers to provide peers
if we can somehow avoid trackers in general, more peers can be found
>>54953215
The torrent protocol is already trackerless. It can fully bootstrap itself via DHT and magnet hashes. The only thing you need is some sort of DHT-based search protocols for finding said magnet links.
(Or just keep on using thin indexing websites and online boards as is the status quo)
>>54951925
you cooking something up in the kitchen right now, mom?
I'd be down for some protocol hacking.
>>54951925
stall protection
more like IPFS
why have the same files in different torrents with different sets of peers when we can just address content by hash making anyone with that file a potential peer
Editing .torrent files or magnet URLs with a private key to keep vidya up to date
Stop the stalled meme
>>54951950
Is that a joke?
>>54951925
be COMPLETELY decentralized
so we could be able to find a file in the whole network just by hash or filename or size
and so each file would be at least 2 times present in a network at one moment
so like when file is uploaded, whole network downloads a bit of it
>>54954039
Why would I be joking?
>>54951925
encryption.
>>54954089
P2P is IP based, you can't hide your ip
>>54954154
Tor, Freenet, i2p etc. are all p2p-based too and yet they're still able to disassociate your IP from the content you're viewing.
>>54954262
It could maybe be an option, but for that to be a default you're going to lose speed because you're juggling data around the network,
>>54951925
The batch idea is intriguing, but it would have to be done in such a way that you would be prompted to add new releases. You wouldn't want random stuff just being tacked on.
It would probably be nice for distributors to not have to make an entirely new torrent when something like subs are updated for a single episode in a batch.
>>54954262
They keep a huge file with a thousand of pieces of random files in your pc, making you a node, and all data is encrypted.
When you request a file you are asking by pieces, and you ask to the node connected to you, if he doesn't have and he keeps asking further until that piece reaches you.
That kind of P2P is self sufficient. You don't need to connect to everyone that downloaded the file you want.
Also Tor isn't 100% P2P, there is servers hosting sites and relays.
Torrent you always rely on someone else that has the file.
is there anything wrong with using the old utorrent releases?
>>54954407
>>54954420
average tripfags
>>54951925
Vuze already has that actually.
>>54951925
>Batch torrents that automatically add new releases.
>Automatic file patching or replacements for v2s.
Both of these can be done with signed content and versioning
>make torrent 2.0
>sign it with privkey
>need to update it because new file/version/better subs/whatever
>new version gets signed and distributed
>peers choose to update (or set auto updating)
They already do this on that torrent website program
>Individual torrent seeder/leecher info.
???? this already exists/please explain further
>>54951936
>bittorrent sync
mission creep
>>54951950
anonymity protocols like I2P and Tor already enable application protocols like torrents and gnutella to operate over them. Building it into the application protocol will just stifle future improvements to mesh net topology.
>>54953850
This so much. private trackers are a cancer, and having 15 torrents that share the same content, but are all dead except for one is dumb as fuck. Content hashable would be nice (but may be a challenge for incorporating with signed content)
>>54954146
no point. Both users know exactly what's transfered and to whom. See mission creep
>>54954262
Tor is neither p2p, nor is it suitable for torrent-like file transfers.
>>54954604
>anonymity protocols like I2P and Tor already enable application protocols like torrents and gnutella to operate over them.
https://blog.torproject.org/blog/bittorrent-over-tor-isnt-good-idea
>>54954904
Good job misreading my post
>Tor is neither p2p, nor is it suitable for torrent-like file transfers.
All I said was that network overlays like I2P and Tor can, as in they are physically compatible with hosting application protocol such as torrents over them
>>54951925
>What features would you want in a Torrent 2.0 protocol?
Network coding. Would allow us to solve the “missing chunk” problem, revive dead torrents, improve swarm efficiency, etc.
Wishlists noted.
Nothing that wasn't already on the list.
11 years into research, experimentation and design here, and getting closer to if not a breakthrough then at least a workable technical compromise.
>>54954948
Do tell us about your research and experimentation
>>54954604
>???? this already exists/please explain further
Meant individual file seeder/leecher info. All you see is info for the whole torrent.
>>54951925
- If it's something like discography, then I'd like it to automatically update itself when new files were added without the need of re-downloading the updated torrent.
- Do some improvements to speed. I don't know, I've never seen my torrents downloading at higher than 10 Mbyte/sec on any torrent client out there - utorrent/transmission/deluge/qbit; but at the same time when I launch my battle-net launcher with peer2peer ON I'm getting download speeds of 14.5-15 Mbyte/sec. How come game launcher is more optimized than popular torrent clients?
- Maybe some layer of protection/encryption. I wan't to see the legit seed|peer statistics, but I'm not happy that anyone can see my IP from their torrent clients, this kind of information should be more private imo.
>>54955075
>I wan't to see the legit seed|peer statistics, but I'm not happy that anyone can see my IP from their torrent clients
...
just think about what you wrote for a moment
less overhead, so that it has more speed, on a roughly 100 kbyte/s upload connection I have to set the upload to about 5 kbyte/s, otherwise my connection becomes clogged.
automatic sharing and searching of trackers. if somebody has a tracker in his client that other don't have already, put the torrent on the new tracker too for everyone. shared trackers for everyone.
streaming functionality a là popcorn, but I know that that is hard to do, I somewhat know how the protocol works
>>54955098
Peer & seed statistics can be available from the tracker. Detailed peer and seed info (IPs) for user clients could be hidden/disabled.
>>54955152
so you want to abolish DHT?
most people want MORE decentralization, you're suggesting we have too much
>>54955075
Torrent:
>I don't want to seed this torrent
or
>I will seed this torrent, but at 10kbps
P2P launcher/patchers:
>Fuck download limit, use all this fucker upload speed, let's save our bandwidth.
>>54955098
>I'm not happy that anyone can see my IP
start here >>54954039
>>54954980
OK.
Cypherpunk. Previously contributed elements to Freenet and I2P. Peripherally, Tor (obfuscation, some countermeasures). Hope you understand if I don't want to say too much: actions speak louder than words.
There's a lot of rabbit holes to chase in this space.
Currently trying to perform a backdoorectomy on Chaum's recent CMix and try to improve it.
Same anon who is developing the messaging platform: same platform. Cover traffic.
Names are particularly hard. Distinctness requires ordering, requires consensus.
I am not Satoshi Nakamoto.
>>54954936
Careful: MSR have a patent on the naïve way. Less naïve ways exist, but many are latency sensitive, not good with bulk traffic. Having some self-synchronising state helps a lot.
>>54954604
you're amazingly fucking stupid if you think tor + torrenting is a good idea
>>54955200
>Tor is neither p2p, nor is it suitable for torrent-like file transfers.
Tor is not p2p, torrent works on it but you will have shit to nothing speeds, so don't use it.
Reading comprehension
>>54955219
who are you quoting? I was just stating a fact
>>54955227
The guy you are quoting to (?)
botnet
>>54955247
this is too confusing nigga
fuck you, fuck him, fuck me, fuck everyone in this thread
>>54955358
Are you quoting right nigga? because the guy you quoted also thinks that torrent on tor is stupid too.
>>54955381
>>54955075
The speed depends on the people who send to you and their actual speeds + limitations if they have set some on uploading
>>54951925
Higher anonymity on par of some latest Japanese P2P soft (not PD)
>>54951925
Higher anonymity on par of some latest Japanese P2P soft (not PD)
>>54954604
things like PD already have Tor built on despite some government already able to crack down cases that use PD
>>54955192
>Careful: MSR have a patent on the naïve way.
Uh oh, more DOA tech that we can't use in practice due to patent trolling?
>>54955192
>>54955947
>Protocol developed for sharing copyrighted media
>Caring about patents
individual file checksumming so duplicate torrent content can be shared among normally unrelated torrents
tracker tracking so similar torrents can be grouped/shared and downed trackers can have their seeds/peers shared similar to bgp route updates
