Hacker News new | past | comments | ask | show | jobs | submit login
Ask HN: What comes after BitTorrent?
64 points by sergiotapia 9 months ago | hide | past | web | favorite | 65 comments
We use to do P2P with Limewire, directly downloading from people's devices.

Next came Soulseek in popularity, directly downloading from people's devices.

Then more than a decade ago, came Bittorrent where you downloaded from the swarm.

What comes next? Are there any next-gen systems being developed to replace bittorrent for file sharing?

BT is still great, and exceptionally good at distributing large, popular files (or groups of files). Why does it need a replacement?

IPFS is probably spiritually closer to BT in terms of censorship and content sharing, and could be a valid replacement if the right incentive based sharing applications are built over it. Right now there is little incentive to seek out and add content to IPFS, if that can be cracked then BT could be wiped out by it.

However, the trend now for most people is to subscribe to spotify, netflix and amazon prime, thinking there is an endless selection of great and original content. Instead you quickly run out of that and end up listening to and watching whatever the algorithm tells you to.

I agree that Netflix has a very limited offering. You won't find any of the classics in there when it comes to movies. You can't really use it as a movie library, in other words.

Spotify actually does have a very large number of songs, and I have no trouble finding classics in many genres - as well as completely new work.

> Spotify actually does have a very large number of songs, and I have no trouble finding classics in many genres

That may be true, but my anecdotal experience shows that I will more often have specific current artists or songs suggested to me than any classics or long-forgotten gems, lending credence to the theory that Spotify cares less about what I'd like and more about what it wants me to listen to

Spotify is still quite limited. To take an example I found recently, it doesn't have Morphine's Like Swimming. The album wasn't a huge success, but still, it's not exactly obscure; it sold 200k copies, made to 67th on Billboard and is owned by UMG.

For me, what often replaces Bittorrent for music is YouTube; still built on copyright infringement, but very unlikely to be blocked or get you in trouble, and it's easier than Spotify (no need for accounts).

I find that Spotify has a lot of mixed albums which are missing tracks due to licensing. I listen to electronic music and a large number of mixed albums are practically useless because of this.

I think the danger is in thinking that streaming providers have everything and are incentivized to show you everything.

e.g. When you pull up an auto-generated playlist or look at recommended suggestions, are you just looking at content which the provider prefers to show you? Can you trust reviews from the same company that's selling you the content? If some content isn't on there, will it just disappear into the abyss? If the provider decides to drop some content, will you ever see it again?

I've had Spotify show me an indie label artist with <1000 listens on Spotify, and <1000 views on Youtube. It was the absolute best thing I've ever seen.

And yes, if content isn't on Spotify for me, it basically doesn't exist. If it's on Bandcamp maybe I will download it.

Spotify allows you to see songs that you have added that have been removed from availability as greyed-out names.

Idk Google All Music Access has a lot of music I otherwise had to torrent for years. Plus what it does not have I can get through YouTube Music. All while being grandfathered into YouTube Red to boot.

> Spotify actually does have a very large number of songs

Large is relative. They have a larger number of Songs, but the actual choice is quite small. It's limited by age and countrys, and even types. They Catalog consists mostly of popular western music and music from the last 10-15 years it seems. Non-Western music or less popular music from the 20th century is quite hard to find on spotify, or any other music-streaming-service. Youtube is a better source for this and they have limitations too. Spotify used to have also audiobooks, but I think they disappeared in the meanwhile because of competing commercial audiobook-streamingservices.

We live in a great world today with all the information and entertainment cheaply available for everyone...as long as you just follow the local mainstream and have not much demand on specific details. It's great and sad at the same time.

Personally I'd rather pay more to Netflix to be able to have a lot more content than using BT or having multiple streaming providers.

The video streaming industry is moving to silos of content with varying levels of success which, regardless of the price, is bad for the user experience.

For example the HBO Go app is bad in terms of usability and (at least in my country) the streaming quality is really bad. I can watch 4K HDR in Netflix, no problem it starts instantly, but HBO Go always looks like some low bandwidth 720p content, specially on camera pans.

I habe been working with IPFS but they have a small "big" problem when working with files in cross browser situations: https://github.com/AquiGorka/test-ipfs-file-sharing

There's always something better. There must be something on the horizon.

"DOS is still great, Why does it need a replacement?"

> There's always something better. There must be something on the horizon.

If something is robust and efficient, why does there need to be something new to replace it? It serves the purpose and solves the problem well, why would you replace it with something new simply for the same of something new?

DOS isn't a great example, DOS is a much bigger system than something like BT. BT is more of a protocol than a piece of software, and even in that case there are piece of software which have existed for decades without the need to be replaced for the reasons above.

New does not mean better, and just because something is old does not mean it's bad or there is something better out there. Perhaps we come up with something that comes after torrents, but we shouldn't do it for the sake of it, it should be because torrents don't work for a use case or we've come up with a faster / more robust way of doing it.

I for one think it's about time we got all our keyboard and mouse replaced by touchscreens and track pads. Newer is better, and things need to be replaced regularly.

Wait, you totally forgot about trackballs. We first need to build a new and shiny trackball into every laptop, and only then replace them all with new and shinny track pads. Let's get that innovation rolling!

Apple is one step ahead of you


This especially does not work when you need a large network effect like Bittorrent does. People know how it works and how to use it and that is why it will be very difficult to overthrow, even if better technology comes along.

It works perfectly fine as it is in my opinion.

And that's why we have new nightmare called USB-C.

The Beaker browser https://beakerbrowser.com/

It uses a P2P protocol called Dat, which is similar to BitTorrent but supports changes to the data and realtime streaming https://www.datprotocol.com/

It's file-sharing mixed into the Web. Beaker uses Dat as a drop-in replacement for HTTP, which means you can browse Dat sites, create new sites within the browser, and use Web APIs to read/write/watch files https://beakerbrowser.com/docs/apis/dat.html

Been playing around with Beaker Browser--super cool project!

So many people I know who are not at all tech savvy buy streaming boxes that offer unlimited free channels. Those things seem to be incredibly popular. This is what has replaced Bittorrent for movie and TV piracy for the masses, or it just enabled piracy when before Bittorrent was just too complex for them. It is now just streamed on demand from a large variety of pirate sites through Kodi and friends.

The people using these generally have no clue they are pirating. It is just this box they bought that offers great value.

This is what has replaced Bittorrent because it is immediately accessible to the masses for $199 or whatever those boxes cost.

What are these “streaming boxes”? I’ve never heard of Kodi either.

Cheap Android boxes running Kodi with a plugin that automatically fetches pirate video streams. Kodi itself is an awesome open source media center that isn't intended to enable piracy, but it supports addons and third parties build plugins that curate pirate streaming sources.

My anecdotal experience is people using an Android device like an Amazon Fire TV Stick that has these sorts of apps sideloaded to watch movies and TV shows. They are sold around $100 with all the necessary setup already completed, allowing people without the ability to connect to an Android device via adb to stream whatever they want.

kodi is extremely popular all over the world. even though the actual product offers no content its addon community is very strong and a lot of piracy addons exist which can be installed easily even if you are not tech savvy.

some addons that are just fronts for piracy sites that host videos but there are some that use Bittorrent to stream content.

getting a box is extremely cheap too. buy any chinese android box and connect it to your tv and you are good to go. but more mainstream there's people who buy boxes, install the addons and resell them at a higher price to people who don't want to mess with technology.

Kodi was formerly XBMC, Xbox Media Center. That homebrew app for the original Xbox.

I think filesharing is mainly just going to die. Streaming is probably the main perpetrator. Hollywood and the music industry are figuring out how to do digital distribution in such a way that consumers don't feel the need for filesharing.

The other is the increasing orwellization of online life, with most computer-mediated interactions and data storage happening in one of a few silos, all of which are cozy with the authorities.

I've made this observation before. File sharing (with the popular BitTorrent and whatever succeeds it) may diminish in use because of streaming, but will not die in the next decade or two. The streaming space is being fragmented further as time passes, and is killing itself. There are many people who're fed up of having to pay and sign up for yet another streaming service and then breaking their head over what content lives where. Unless the providers get their act together by letting go of the greed to be in control and the greed to make a lot of money — both with fragmentation as the solution — the scene from the last few years and continuing for the next few years only acts as a fuel for file sharing. One doesn't have to think about which provider has what content, with ads or without ads and all that stuff. Just go to your favorite file sharing site and get what you want, when you want and how you want.

For video content as well as content that's not massively popular in the population, as it was before, file sharing still provides the best experience and the best choices (content, bandwidth requirement, file sizes, file formats, etc.) for those who know how to use it.

File sharing existed with the dawn of personal computing, it will exist until it dies.

Torrenting too. Getting all the adobe products I need for $35/mo is a perfect fit.

Adobe pricing hurts a ton when you are a casual user, though.

Most subscription apps or products are way too expensive for casual users (those who don't make money from using the apps or products). It just starts piling up and could become enormous - one subscription for this, one for that, one for another and so on.

I'm just not thrilled about the idea of paying rent for software. It's hard living without Adobe, but so far, I'm managing.

I have daily withdrawals from Illustrator, though.

Wish they had another pricing model. I use Photoshop a few times a year. Would happily pay a little to use it for a few days, and then let it sit cold on my disk for a few months before starting it up again for another small fee.

Bittorrent is almost two eras here, regular torrent files and magnet links with their associated DHT. Splitting off the metadata to a giant DHT, is a huge (albeit backwards compatible) change.

I think that shows why Bittorrent might be here to stay for a while, it's willing to make pretty deep changes to stay with the times.

WebTorrent (https://webtorrent.io) is the first torrent client that works in the browser.

It's written completely in JavaScript – the language of the web – and uses WebRTC for true peer-to-peer transport. No browser plugin, extension, or installation is required. The WebTorrent protocol works just like BitTorrent protocol, except it uses WebRTC instead of TCP/uTP as the transport protocol.

Using open web standards, WebTorrent connects website users together to form a distributed, decentralized browser-to-browser network for efficient file transfer.

There's no real incentive for regular BitTorrent clients to spend the development time to add support, because the only ones with something to gain are WebTorrent users.

Existing BitTorrent clients, on the other have something to lose:

* WebTorrent peers are less likely to be connectible (behind NAT etc.)

* WebTorrent peers are more likely to "hit and run" (by navigating away from the page, you stop seeding)

* WebTorrent peers are more likely to favor sequential downloads instead of rarest first (for in-browser playback)

Being incompatible with regular BitTorrent means it does not get any utility from existing swarms (unless regular BitTorrent clients gain widespread support; unlikely).

This means WebTorrent is mostly relevant for content delivery, for use cases a site offloads bandwidth usage onto their uses (like PeerTube, for example), and the site itself always runs a seed.

Mostly true, except with BitTorrent you get tricks for NAT traversal that WebRTC will never have and failure is an option. It works without a STUN/TURN central server, as long as the swarm is sufficiently connected.

Why is this good?

The music and film industries mostly ;-)

But jokes aside, I don't see much need for something to replace BitTorrent as a mean of distributing content. Gradual improvements sure, something completely new not really.

Where big changes are going to be needed is content discovery - as big torrent search sites disappear it's becoming more apparent they and not the trackers are the weak point.

Online payment is so much easier, and incredibly more accepted than it was during the wild Napster/Limewire/Torrent days. Companies are finally figuring out that if they just make things available, people will pay for it, too.

Network speeds have increased sufficiently that keeping vast local libraries of content the way we used to makes less sense than streaming; I get speeds from Netflix or Amazon comparable or better than I used to get over USB from external hard disks.

unfortunately companies are still far behind with availability. evident from comparing netflix catalog from different countries. additionally a lot of older more obscure content sometimes can only be found via non legal methods. this is unfortunate but true.

The two things I don't see bittorrent utilized for nowadays are livestreaming, and file matching. Sure, there's some demos of bittorrent live, but it's not hit the mainstream yet. 30-60 second stream delay should be plenty of time to sync video to everybody. File matching would be neat too. Just scan your files against a DHT database and you'll never have to worry about losing them again. Just as long as there's seeds.

There's usenet, and importantly another layer, plex. the future is some people downloading everything and all their friends using their plex

I honestly think that the idea of letting people search inside a torrent before fully downloading it has been overlooked and might be the next big step in P2P file-sharing: https://github.com/lmatteis/torrent-net

The idea is that you can actually query stuff inside a torrent; like actually perform an SQL query and only the pieces relevant to the query will be downloaded.

This sort of allows for distributed querying if you think about it, since your query could be satisfied by many different peers that are seeding/leeching.

Imagine a Web built this way. Where sites are served by people that visit them and are not just static sites, but fully queryable as you'd expect in the normal HTTP web.

I wrote more on this here: https://medium.com/@lmatteis/torrentnet-bd4f6dab15e4

Along the same lines as IPFS—and having some similar goals [0]—is the DAT project and related protocols.

[0]: https://github.com/datproject/docs/blob/master/docs/faq.md#h...

There's torrent via I2P, which provides a different anonymity-vs-performance tradeoff compared to normal BitTorrent.

I would say that BitTorrent just fits is particular design constraints fairly well, so I don't see anything replacing it without changing the use case patterns.

People are always trying out new protocol features and enhancements. Clients ship with the ones that have turned out to be the most useful.

In other words, BitTorrent will be the next BitTorrent.

Compare modern clients that can easily handle seeding tens of thousands of files across hundreds of different trackers with minimal resource utilization with the reference client people used back in the early 2000s. You only had one way to connect to trackers and other clients, the traffic was trivial to detect and filter, and the client only support one file at a time along with using 100% of a cpu.

Implement the thing they are attempting to make the in the comedy TV series Silicon Valley and have a torrent style decentralized internet. Except use computers as well as phones.

To each network their pros and cons, I think if you look at it from a pirating point of view, torrents seems to tick all the technical boxes.

If you talk about it in terms of corporate use / pirating fads / user experience etc. I don't know At this point the success is not technical anymore and will depend on the network effect (pun non intended).

I bet the next Big stuff is Enterprise P2P.

Some content providers are using P2P to complement their CDNs with more offnet capilarity (cannot tell details in public).

Others like Alibaba are using P2P to syncronize content among servers. Thing of it as a faster and simpler way to have server farm contents un sync.

What I'm missing yet is using P2P for configuration, and backup.

@eb0la is probably right. The trends and challenges in the Enterprise favor p2p solutions to solve several different problem vectors.

The first is Scale. Ever greater scale in terms of data, number of endpoints, workflows, locations… you name it. Moore’s law means more computers doing more things at every point in the Enterprise, but especially at the edge. Scaling enterprise systems will be one of the great challenges over the next decade and not just for the web monsters. Everyone will need to contend with scale. The good news is nothing scales like P2P. It’s organic. When every client is also a server, greater scale simply means even more supply and faster speeds. It’s the reason 20 engineers at bittorrent could build software that moves whole percentage points of total Internet traffic (link to recent exabyte blog post?), traffic volume ten or twenty times the size of even the largest websites with thousands of employees and large infrastructure and operating budgets.

The second is Reliability. Making workflows reliable is quite expensive with traditional architectures. Building in redundancy for high availability can be very expensive and grossly inefficient in the client-server model. Conversely, P2P is naturally resilient and reliable when there are many potential sources for the data. If one is not available, no problem, pick among the thousands of alternatives. P2P is resilient while remaining efficient in resource utilization.

Finally, as mentioned above, there is a growing need to operate at the edge of the network, where resources are spread over a wide area and potentially over low-quality networks. If the experts are right, there will soon be 10 times the data volume at the edge as in the cloud. (see gartner). Managing workflows at the edge of the network will require a centrally managed P2P solution (like Resilio) utilizing a combination of technologies to solve all of the above.

Managing Enterprise workflows reliably and at scale, in the cloud or on the edge, favors a p2p solution and you already see it in various IoT, container orchestration and edge computing architectures. Expect to see more.

BitTorrent spin-off Resilio is doing enterprise P2P. It is the next big thing.

The problem with Bittorrent is that it is not good at lots of individual small files. This is where Google Drive and Dropbox excel (mostly.)

When there are lots of small files, you usually want all of them, something that a torrent guarantees and file by file downloads do not.

I'm aware that both Google Drive and Dropbox can create zip archives of whole folders, but who guarantees that the content hasn't been deleted or replaced? Integrity is more important than convenience.

What sort of issues do you have with lots of individual files? Are you talking about bundling them under torrents and dealing with that? Or having trackers handle +1000s of torrents?

Is there a self-hosted solution that does? Like IPFS or Dat?

If it comes to prevention of piracy, I think a lot of torrent sites go underground, behind login or on Tor network.

What comes next? It must be Netflix, given the decrease in BitTorrent seeders over the last decade or so.

I may have some anecdotal data to back this claim that BitTorrent seeders have decreased in the last decade (or more like the last half a decade), but there are plenty of seeders on hundreds, if not thousands, of private sites and trackers. I don't think that has decreased a lot. Maybe someone who has studied this better to state how things have changed.

IPFS comes to mind.

But I doubt anything will displace BitTorrent anytime soon. It's quiet entranched at this point

Seems like IPFS potentially?

IPFS looks very hopeful.

The documentation is very light on for now though. There are some large organizations that use NFS or AFS in large networks, IPFS looks like a promising replacement, but I think they would need more documentation and stability before considering a switch.

Netflix, HBO, Amazon, and Apple providing music and video at a reasonable price, high quality, and easy to use, while respecting their customers in a way the RIAA and MPAA never did

Applications are open for YC Summer 2019

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact