Most Torrent Sites Blocked in India, about frostwire HOT 20 OPEN

The Other option is using firebase remote config.

from frostwire.

we have already a way to do remote config.
we're not in india though, we have no clue what search engines work there, perhaps you want to become a contributor and integrate new search performers

from frostwire.

In India the ISPs block the domain names, so there are two workarounds first one is using proxies to access all the search engines and download torrent files/get the magnet link and then download the file via normal network so the proxy won't have much load apart from being used for searching or getting magnet link. Second option is using remote config method and I will try to update you guys with the latest working mirror sites, so it keeps on working.

from frostwire.

one important caveat with proxies, they can't be behind services like cloudfare, we don't have the means to defeat their anti-dos/captcha-like protections, we've had to remove several search engines that decided to use their anti-ddos tech

from frostwire.

Then there is only one way left and that is to use remote config.

from frostwire.

there's an ultimate way left, p2p search, impossible to block.

from frostwire.

Yeah, but is there any plan from you guys to add it?

from frostwire.

it's a long plan that requires a few building blocks.

Would love to experiment with a design inspired in a (sharded) blockchain that indexes torrent metadata and the sha1 hashes and a gossip based search protocol.

Records in the blockchain are as simple as:
torrent name, size, sha1 hash
optionally include the names and sizes of the files indexed by the torrents.

Clients are constantly downloading the last blocks of the blockchain shard they're part of and creating a local index.

When you perform a search, you can search first your local shard index, but you can also talk to a few peers and ask them for what they know.

As clients download new torrents, they can propose new torrents to be indexed, here based on some reputation heuristics, spam checks, a set of reputable nodes can decide to add new records or not to the blockchain shards.

If spam becomes an issue, then we can introduce a token that has to be spent (perhaps a new token, or a token from an existing blockchain with real economic value) so that in order to post you have to spend a resource.

If pulled off, this could potentially put out of business (or decimate at least) sketchy torrent index sites, just blockchain that data.

from frostwire.

This will take a lot of time. Is there any temporary work around until this method is implemented?

from frostwire.

any temporary work around only delays working on the real solution.

from frostwire.

perhaps an in between (simpler) solution will be to proxy searches to other peers that are not blocked.

from frostwire.

Yeah, it would be great.

from frostwire.

In India the ISPs block the domain names

@fahimanwer Do they only block the domains? If you use a DNS like OpenDNS, can you directly access those websites?

@gubatron What if Frostwire incorporated an entire DHT crawler like BTDigg or Magnetico? Every user would host its own database and the only connection needed would be the connection to the DHT itself. It isn't something to be "on" by default, but for those who need it, it is the most decentralized search that one can get.

from frostwire.

Thank you for bringing these ideas up, I wasn't aware of these crawlers and we're writing down the roadmap for FrostWire 7.

The 3 big tickets are:

• Bittorrent 2.0 compatibility
• IPFS integration
• Distributed search, DHT crawling is one of the possibilities, but we're thinking of something even cooler.

from frostwire.

• stributed search, DH

I tried using both OpenDNS as well as Cloudflare DNS or even Googles but nothing works, sometimes VPNs do but that is also not permanent.

from frostwire.

• Bittorrent 2.0 compatibility

Seems very interesting, till when can we expect it? The best possible feature for me personally is to be able to evade using VPNs to search for torrent.

from frostwire.

• Distributed search, DHT crawling is one of the possibilities, but we're thinking of something even cooler.

@gubatron You didn't you say what else you have in mind, but here is a brute-force way to achieve a partially decentralized search without the need to - at least for now - incorporate an entire DHT crawler into every Frostwire build: self-host one of those DHT crawlers, crawl the DHT (maybe) everyday and then dump the database in a folder on a self-hosted IPFS node. The folder should/could have only 5 files: a file for a database with torrents found in the last 24 hours, another one for those found in the last 72 hours, another one for those found in the last 7 days, another one (well, probably more than one...) with all the torrents ever found in the DHT, and an index file pointing to those other files. Register an IPNS for such a folder and hardcode it in Frostwire so that it looks out for that index file. If the user has a local IPFS node running then it will be able to download those databases and have a full DHT search. This way, the only native code to be written for the crawler (but not really for the crawler) will be the one to search the database on the user-side. The index file and the database files could be automatically imported to the user's local IPFS node.

@fahimanwer Can you access in the browser when using DNS-over-HTTPS?

from frostwire.

self-host one of those DHT crawlers

That's not an option, it's centralization, single point of failure, and also a liability.

Rather have every node do a little bit of work towards a bigger dataset in completely descentralized fashion.

It can be done with IPFS, but it can also be done with jlibtorrent and torrents.

from frostwire.