8 ms·

Distributed web not ready for Runet cutoff from the Internet

85 points by agentofuser 8y ago

Gaelan 8y ago
Oh lord, what’s this going to do to the blockchains?
- sgeisler 8y ago
  At least bitcoin will be mostly unaffected since you can receive blocks via satellite [1] and if even one satellite connected node is inside Runet then you can at least keep up with the blockchain. Sending transactions can be done via sms [2]. The main problem would be mining in Russia, but if miners know about the shutdown in advance and it's economically feasible then they can get satellite internet to upload found blocks. [1] https://blockstream.com/satellite/ https://blockstream.com/satellite/ [2] https://twitter.com/SamouraiWallet/status/954271930365947904 https://twitter.com/SamouraiWallet/status/954271930365947904
  - ForHackernews 8y ago
    is anyone actually using this?
    - wongarsu 8y ago
      One person is enough. If one person inside Runet receives blocks from the wider internet (via satellite or other means), their client will automatically forward the blocks to all clients they are connected to, quickly spreading them accross Russia. And because Bitcoin follows the chain of blocks that took the most work to create, latency isn't an issue. Mining (and making a profit while doing so) is more latency sensitive, but should be able to tollerate a satelite uplink. And again, one person having a satelite uplink is enough to give all Russian miners a chance.
      - ForHackernews 8y ago
        Is there even much mining in Russia? My understanding is the overwhelming majority of bitcoin mining is in China.
        wongarsu 8y ago
        Mining is mostly a game of getting cheap electricity, and it's hard to beat China (where infrastructure construction is often about creating jobs and projecting power instead of classic profitability). It seems like electricity prices in Russia are similar to the US. The legal position of bitcoin seems to be worse in Russia though, so you would expect less mining than in the US, and a lot less than in China.
marknadal 8y ago
This is total FUD. The Decentralized Web is a lot bigger than just only the 2 projects they mention. Developers in our community ( https://github.com/amark/gun https://github.com/amark/gun ) are already running local multi-cast discovery (working!) and have Bluetooth in the works. And we can do extended peer discovery through decentralized signaling relays (any/all peers contribute to this, WebRTC, browser, or other). Internet Archive is using us for their dweb mirror. Additionally the system uses daisy-chain mesh net algorithms to proxy messages regardless of whether you are directly connected to peers or not. Bring it, Russia.
pfraze 8y ago
I can attest that the blog post is accurate, at least for the Dat protocol. We used the BitTorrent mainnet DHT for a while, but results were so bad that we retired it and fell back to our tracker. We're working on the Hyperswarm DHT now to get things more resilient. We're taking up Daniel's other action points as well.
- gpm 8y ago
  I'm curious, what's wrong with the BitTorrent DHT that you can improve in your own?
zzzcpan 8y ago
The sad reality of modern "decentralized" software is that organizations, people behind it don't even have any intentions to make it decentralized, to give up control, like BitTorrent did. Nobody cares about the main thing that decentralization is all about - making rule/law enforcement on others practically impossible. There was a great series of posts about decentralization from BitTorrent people recently: https://medium.com/@simonhmorris/why-bittorrent-mattered-bittorrent-lessons-for-crypto-1-of-4-fa3c6fcef488 https://medium.com/@simonhmorris/why-bittorrent-mattered-bit... https://medium.com/@simonhmorris/if-youre-not-breaking-rules-you-re-doing-it-wrong-bittorrent-lessons-for-crypto-2-of-4-72c68227fe69 https://medium.com/@simonhmorris/if-youre-not-breaking-rules... https://medium.com/@simonhmorris/intent-complexity-and-the-governance-paradox-bittorrent-lessons-for-crypto-3-of-4-1d14ac390f3f https://medium.com/@simonhmorris/intent-complexity-and-the-g... https://medium.com/@simonhmorris/decentralized-disruption-who-dares-wins-bittorrent-lessons-for-crypto-4-of-4-f022e8641c1a https://medium.com/@simonhmorris/decentralized-disruption-wh...
- fenwick67 8y ago
  > ...organizations, people behind it don't even have any intentions to make it decentralized, to give up control, like BitTorrent did Huh? Where is this coming from? Can the IPFS or DAT community leaders reign anything in? There are bootstrap nodes but I'm pretty sure these are just practical stopgaps for now, not a means of control.
- colordrops 8y ago
  Bitcoin?
  - Etheryte 8y ago
    While in theory Bitcoin is decentralized, in practice roughly half of the coins in existence are from a single mining company and 74% of global mining power currently resides in China, belonging to a handful of owners. https://bitcoinmagazine.com/articles/report-links-74-bitcoin-mining-china-sees-threat-network/ https://bitcoinmagazine.com/articles/report-links-74-bitcoin...
ezoe 8y ago
The last time I checked IPFS is a joke. There are only two implementations of IPFS protocol and both come from the same group of developers. On the other hand, there are so many BitTorrent protocol implementations from totally independent group of developers. Also, the namespace system of IPFS is just beyond stupid. They openly recommend relying the traditional central DNS to resolve the human readable name to hash. I didn't bother to examine the Dat, but it looks like worse than IPFS.
- slaymaker1907 8y ago
  Is server side JS a joke because there is basically one implementation, Node? Number of independent implementations is a mostly useless metric in the FOSS world. * it is not useless for proprietary software since then reliance on a single vendor is a much more serious issue.
- rklaehn 8y ago
  There is a version of libp2p written in rust, developed by parity.io: https://github.com/libp2p/rust-libp2p https://github.com/libp2p/rust-libp2p . IPFS is much younger than bittorrent, so it is not surprising that there are not as many implementations. I am not sure why that makes it "a joke". All that is needed for IPFS to function in a russia-only internet are a few russian bootstrap nodes. Bootstrap nodes are not in any way special, so any node can serve as a bootstrap node.
- anticensor 8y ago
  > namespace system of IPFS Namespace companion of IPFS is called IPNS.
ytjohn 8y ago
One of them that will, or at least should be able to continue, is scuttlebutt. Scuttlebutt clients on an isolated LAN can share messages with each other. To go beyond the LAN, you point at a "pub server", which anyone can host anywhere. Yes, only russian-based pub servers will be available, and that assumes they are using russian based DNS to point to them (or have a known static IP). However, for the test of communicating within your region, scuttlebutt will hold up. The protocol, clients, and pub servers (which are really just automated clients) are designed with an offline-first approach, then local, then regional, then global. Messages are copied between clients as they come into contact with each other. One way to visualize it is by picturing islands of people gossiping locally, and then having ships carry that scuttlebutt from island to island. https://www.scuttlebutt.nz/ https://www.scuttlebutt.nz/
- gregknicholson 8y ago
  > To go beyond the LAN, you point at a "pub server", which anyone can host anywhere. You could also get messages into and out of Russia by physically moving a laptop or phone with a Scuttlebutt client. Forgive my ignorance here but, supposing there was a mesh network spanning across the block, would Tor, cjdns, Hyperboria or Yggdrasil be low-level enough to get around a block like this?
  - neuraxon77 8y ago
    Yes, any network link to an outside pub would eventually sync that pub to the rest of the network. This makes Secure Scuttlebutt a [delay tolerant network](https://en.wikipedia.org/wiki/Delay-tolerant_networking https://en.wikipedia.org/wiki/Delay-tolerant_networking).
- state_less 8y ago
  Isn't the design motivated in part because one (or more) of the key contributors actually are/were sailors? https://www.theatlantic.com/technology/archive/2017/05/meet-the-counterantidisintermediationists/527553/ https://www.theatlantic.com/technology/archive/2017/05/meet-...
hcs 8y ago
Hm, I had thought that go-ipfs saved some swarm peers between sessions, but reading over the docs now I see no evidence of this.
- d2wa 8y ago
  There is no implementation of this in their code either.
lightgreen 8y ago
> Russia is preparing a nation-wide experiment where the whole country temporarily disconnects from the global Internet to see if the country can rely on Runet alone I think this is lie, fake news and misinformation. I tried to find the source of it. Lots of site quote RBC website (more or less respectful news agency in Russia). The closest article on this subject is this: https://www.rbc.ru/technology_and_media/08/02/2019/5c5c51069a7947bef4503927 https://www.rbc.ru/technology_and_media/08/02/2019/5c5c51069... It discusses this topic but doesn’t mention experiments or temporary disconnect. Could be redacted though.
- lstodd 8y ago
  Why, it does mention them. > Therefore, they came to the conclusion that market participants need to organize exercises or something similar in order to understand how all this can be implemented in practice. Edit: also this https://www.rbc.ru/technology_and_media/15/02/2019/5c666af39a79478787a38cd3 https://www.rbc.ru/technology_and_media/15/02/2019/5c666af39...
  - lightgreen 8y ago
    There’s huge difference between some experiments on network equipment which doesn’t affect network connectivity and “nation-wide experiment where the whole country temporarily disconnects from the global Internet”.
- PavlovsCat 8y ago
  recent discussion: https://news.ycombinator.com/item?id=19135085 https://news.ycombinator.com/item?id=19135085 https://news.ycombinator.com/item?id=19132665 https://news.ycombinator.com/item?id=19132665
- ascii_only 8y ago
  I don't think you are being honest in your attempt to search. It takes no time to find reliable sources for this. Here is for example article from Tass: https://tass.ru/obschestvo/5914675 https://tass.ru/obschestvo/5914675
sodosopa 8y ago
Putting on my conspiracy hat, blocking the global web from within Russia would also make it easier to have huge sweeping "purges" toward targeted groups, without Western interference.

85 points by agentofuser 8y ago

Gaelan 8y ago

Oh lord, what’s this going to do to the blockchains?

sgeisler 8y ago
At least bitcoin will be mostly unaffected since you can receive blocks via satellite [1] and if even one satellite connected node is inside Runet then you can at least keep up with the blockchain. Sending transactions can be done via sms [2]. The main problem would be mining in Russia, but if miners know about the shutdown in advance and it's economically feasible then they can get satellite internet to upload found blocks. [1] https://blockstream.com/satellite/ https://blockstream.com/satellite/ [2] https://twitter.com/SamouraiWallet/status/954271930365947904 https://twitter.com/SamouraiWallet/status/954271930365947904
- ForHackernews 8y ago
  is anyone actually using this?
  - wongarsu 8y ago
    One person is enough. If one person inside Runet receives blocks from the wider internet (via satellite or other means), their client will automatically forward the blocks to all clients they are connected to, quickly spreading them accross Russia. And because Bitcoin follows the chain of blocks that took the most work to create, latency isn't an issue. Mining (and making a profit while doing so) is more latency sensitive, but should be able to tollerate a satelite uplink. And again, one person having a satelite uplink is enough to give all Russian miners a chance.
    - ForHackernews 8y ago
      Is there even much mining in Russia? My understanding is the overwhelming majority of bitcoin mining is in China.
      - wongarsu 8y ago
        Mining is mostly a game of getting cheap electricity, and it's hard to beat China (where infrastructure construction is often about creating jobs and projecting power instead of classic profitability). It seems like electricity prices in Russia are similar to the US. The legal position of bitcoin seems to be worse in Russia though, so you would expect less mining than in the US, and a lot less than in China.

marknadal 8y ago

This is total FUD. The Decentralized Web is a lot bigger than just only the 2 projects they mention. Developers in our community ( https://github.com/amark/gun https://github.com/amark/gun ) are already running local multi-cast discovery (working!) and have Bluetooth in the works. And we can do extended peer discovery through decentralized signaling relays (any/all peers contribute to this, WebRTC, browser, or other). Internet Archive is using us for their dweb mirror. Additionally the system uses daisy-chain mesh net algorithms to proxy messages regardless of whether you are directly connected to peers or not. Bring it, Russia.

pfraze 8y ago

I can attest that the blog post is accurate, at least for the Dat protocol. We used the BitTorrent mainnet DHT for a while, but results were so bad that we retired it and fell back to our tracker. We're working on the Hyperswarm DHT now to get things more resilient. We're taking up Daniel's other action points as well.

gpm 8y ago
I'm curious, what's wrong with the BitTorrent DHT that you can improve in your own?

zzzcpan 8y ago

The sad reality of modern "decentralized" software is that organizations, people behind it don't even have any intentions to make it decentralized, to give up control, like BitTorrent did. Nobody cares about the main thing that decentralization is all about - making rule/law enforcement on others practically impossible. There was a great series of posts about decentralization from BitTorrent people recently: https://medium.com/@simonhmorris/why-bittorrent-mattered-bittorrent-lessons-for-crypto-1-of-4-fa3c6fcef488 https://medium.com/@simonhmorris/why-bittorrent-mattered-bit... https://medium.com/@simonhmorris/if-youre-not-breaking-rules-you-re-doing-it-wrong-bittorrent-lessons-for-crypto-2-of-4-72c68227fe69 https://medium.com/@simonhmorris/if-youre-not-breaking-rules... https://medium.com/@simonhmorris/intent-complexity-and-the-governance-paradox-bittorrent-lessons-for-crypto-3-of-4-1d14ac390f3f https://medium.com/@simonhmorris/intent-complexity-and-the-g... https://medium.com/@simonhmorris/decentralized-disruption-who-dares-wins-bittorrent-lessons-for-crypto-4-of-4-f022e8641c1a https://medium.com/@simonhmorris/decentralized-disruption-wh...

fenwick67 8y ago
> ...organizations, people behind it don't even have any intentions to make it decentralized, to give up control, like BitTorrent did Huh? Where is this coming from? Can the IPFS or DAT community leaders reign anything in? There are bootstrap nodes but I'm pretty sure these are just practical stopgaps for now, not a means of control.
colordrops 8y ago
Bitcoin?
- Etheryte 8y ago
  While in theory Bitcoin is decentralized, in practice roughly half of the coins in existence are from a single mining company and 74% of global mining power currently resides in China, belonging to a handful of owners. https://bitcoinmagazine.com/articles/report-links-74-bitcoin-mining-china-sees-threat-network/ https://bitcoinmagazine.com/articles/report-links-74-bitcoin...

ezoe 8y ago

The last time I checked IPFS is a joke. There are only two implementations of IPFS protocol and both come from the same group of developers. On the other hand, there are so many BitTorrent protocol implementations from totally independent group of developers. Also, the namespace system of IPFS is just beyond stupid. They openly recommend relying the traditional central DNS to resolve the human readable name to hash. I didn't bother to examine the Dat, but it looks like worse than IPFS.

slaymaker1907 8y ago
Is server side JS a joke because there is basically one implementation, Node? Number of independent implementations is a mostly useless metric in the FOSS world. * it is not useless for proprietary software since then reliance on a single vendor is a much more serious issue.
rklaehn 8y ago
There is a version of libp2p written in rust, developed by parity.io: https://github.com/libp2p/rust-libp2p https://github.com/libp2p/rust-libp2p . IPFS is much younger than bittorrent, so it is not surprising that there are not as many implementations. I am not sure why that makes it "a joke". All that is needed for IPFS to function in a russia-only internet are a few russian bootstrap nodes. Bootstrap nodes are not in any way special, so any node can serve as a bootstrap node.
anticensor 8y ago
> namespace system of IPFS Namespace companion of IPFS is called IPNS.

ytjohn 8y ago

One of them that will, or at least should be able to continue, is scuttlebutt. Scuttlebutt clients on an isolated LAN can share messages with each other. To go beyond the LAN, you point at a "pub server", which anyone can host anywhere. Yes, only russian-based pub servers will be available, and that assumes they are using russian based DNS to point to them (or have a known static IP). However, for the test of communicating within your region, scuttlebutt will hold up. The protocol, clients, and pub servers (which are really just automated clients) are designed with an offline-first approach, then local, then regional, then global. Messages are copied between clients as they come into contact with each other. One way to visualize it is by picturing islands of people gossiping locally, and then having ships carry that scuttlebutt from island to island. https://www.scuttlebutt.nz/ https://www.scuttlebutt.nz/

gregknicholson 8y ago
> To go beyond the LAN, you point at a "pub server", which anyone can host anywhere. You could also get messages into and out of Russia by physically moving a laptop or phone with a Scuttlebutt client. Forgive my ignorance here but, supposing there was a mesh network spanning across the block, would Tor, cjdns, Hyperboria or Yggdrasil be low-level enough to get around a block like this?
- neuraxon77 8y ago
  Yes, any network link to an outside pub would eventually sync that pub to the rest of the network. This makes Secure Scuttlebutt a [delay tolerant network](https://en.wikipedia.org/wiki/Delay-tolerant_networking https://en.wikipedia.org/wiki/Delay-tolerant_networking).
state_less 8y ago
Isn't the design motivated in part because one (or more) of the key contributors actually are/were sailors? https://www.theatlantic.com/technology/archive/2017/05/meet-the-counterantidisintermediationists/527553/ https://www.theatlantic.com/technology/archive/2017/05/meet-...

hcs 8y ago

Hm, I had thought that go-ipfs saved some swarm peers between sessions, but reading over the docs now I see no evidence of this.

d2wa 8y ago
There is no implementation of this in their code either.

lightgreen 8y ago

> Russia is preparing a nation-wide experiment where the whole country temporarily disconnects from the global Internet to see if the country can rely on Runet alone I think this is lie, fake news and misinformation. I tried to find the source of it. Lots of site quote RBC website (more or less respectful news agency in Russia). The closest article on this subject is this: https://www.rbc.ru/technology_and_media/08/02/2019/5c5c51069a7947bef4503927 https://www.rbc.ru/technology_and_media/08/02/2019/5c5c51069... It discusses this topic but doesn’t mention experiments or temporary disconnect. Could be redacted though.

lstodd 8y ago
Why, it does mention them. > Therefore, they came to the conclusion that market participants need to organize exercises or something similar in order to understand how all this can be implemented in practice. Edit: also this https://www.rbc.ru/technology_and_media/15/02/2019/5c666af39a79478787a38cd3 https://www.rbc.ru/technology_and_media/15/02/2019/5c666af39...
- lightgreen 8y ago
  There’s huge difference between some experiments on network equipment which doesn’t affect network connectivity and “nation-wide experiment where the whole country temporarily disconnects from the global Internet”.
PavlovsCat 8y ago
recent discussion: https://news.ycombinator.com/item?id=19135085 https://news.ycombinator.com/item?id=19135085 https://news.ycombinator.com/item?id=19132665 https://news.ycombinator.com/item?id=19132665
ascii_only 8y ago
I don't think you are being honest in your attempt to search. It takes no time to find reliable sources for this. Here is for example article from Tass: https://tass.ru/obschestvo/5914675 https://tass.ru/obschestvo/5914675

sodosopa 8y ago

Putting on my conspiracy hat, blocking the global web from within Russia would also make it easier to have huge sweeping "purges" toward targeted groups, without Western interference.