9 ms·

Open Libernet: a Bitcoin-based fully encrypted mesh networking protocol

126 points by gasull 13y ago

jsaxton86 13y ago
A large scale mesh network is interesting to me for a couple of reasons: * Consumers would no longer be dependent on monopolies like Comcast for internet service (OSI layer 1) * It is much more resistant to eavesdropping and censorship (OSI layers 2-3) These seem like two totally different problems to me, and I'm still struggling to understand how the physical layer part can work on a large scale. I get how they could work on a small scale. For example, you could easily set up an ad-hoc wifi network in a college dorm for, say, file sharing purposes. But in most other settings, I don't get how it could work. Most people don't even talk to their neighbors, so even if you lived in a neighborhood that had a high enough density of nodes for this to work, why would you want to share a little ad-hoc network with your neighbors? Of course, wireless networks alone aren't going to be sufficient to build a large-scale mesh network that could disrupt existing ISPs. The latency alone makes this infeasible. So you need someone to invest in physical links, especially physical links that cover long distances. That's why the BTC part of this proposal is so great -- it provides a way to compensate these people. Of course, this is just a slightly different problem than the one we have today, but instead of being stuck with Comcast we're stuck with whoever runs these physical links.
- lotsofmangos 13y ago
  wireless networks alone aren't going to be sufficient to build a large-scale mesh network that could disrupt existing ISPs. The latency alone makes this infeasible. I'm not so sure, your links are almost c, whereas optic fiber is substantially slower, so it depends entirely on how little processing you can get away with in the message passing code. edit - was just thinking about this, an arm cortex runs at ~ 2,000 MIPS, which gives you just under 15 cm per instruction for radio in air. Fiber is about a third slower, so your links do not actually have to be massively long to give you quite a few cycles to play with.
  - jsaxton86 13y ago
    I think it mostly depends upon the range of each of the nodes. If your nodes have an average range of 820 feet (the range of 802.11n (in good weather, I assume)), and you have 1 node exactly every 820 feet between NY and SF, that's over 16,000 links. If each link has 10ms of latency, that's 16s of latency.
    - ZenoArrow 13y ago
      16s of latency is interesting to think about, do think we'll get around this issue with local web page caching (obviously won't work for non turn-based games). SyncNet or something similar could prove very useful... http://jack.minardi.org/software/syncnet-a-decentralized-web-browser/ http://jack.minardi.org/software/syncnet-a-decentralized-web...
  - ris 13y ago
    It's got nothing to do with the propagation speed. Wireless has high latency because it requires comparatively high amounts of processing, has to operate over a shared channel (requiring collision avoidance), is subject to unexpected interference...
    - adestefan 13y ago
      You also need to send a lot of additional information for error correction.
- JoshTriplett 13y ago
  > It is much more resistant to eavesdropping and censorship (OSI layers 2-3) Not inherently. Arguably, a large-scale mesh network would be more susceptible to interference, due to the larger number of distinct parties involved in each communication. The difference is that because a mesh network assumes the lack of trust in the intermediate nodes, a sensible design for one has to include end-to-end security.
- rmrfrmrf 13y ago
  I think you have your OSI layers mixed up.
- jewel 13y ago
  > Most people don't even talk to their neighbors, so even if you lived in a neighborhood that had a high enough density of nodes for this to work, why would you want to share a little ad-hoc network with your neighbors? I've been thinking about this recently. The idea is that it's transparent. Imagine, for example, if Apple started shipping this on all iOS devices, or it became a common android feature. Here is a blog post I wrote on the topic last month: http://stevenjewel.com/2014/01/android-mesh/ http://stevenjewel.com/2014/01/android-mesh/
tinco 13y ago
Has no reference implementation.
- cmpaul 13y ago
  "20. What is the status of Open Libernet at this time? We are still at idea stage. We’ve built a few simulations to test some parts of the system, but the main development hasn’t begun. We have published our whitepaper, which explains – to some degree of detail – the proposed protocol, and we’re still waiting for review. For that, if you happen to have any insight on routing protocols, networking or software development that you would like to share with us, please do so. If you think something with the protocol is not right or needs to be improved, or if you have a better idea on how to do something, also do not hesitate to voice your opinion. This is a community project and we are more than happy to receive all the help we could get. Also, everything we’re doing is Open Source, including the white-paper, the concept and in the future the code. You are free to take anything you like, modify, publish and release, then tell us about it. Or if you prefer, don’t."
deanclatworthy 13y ago
I read these mesh network posts with great interest. I really like the idea of a neighbourhood having their own private network for communication. Sadly this project is not yet ready. Let's say I found 10 people in my neighbourhood with 10 routers what's out there that's easy to set up that would get us started with a little mesh network?
synctext 13y ago
This is already moving within IETF: https://datatracker.ietf.org/doc/draft-pouwelse-perpass-shadow-internet/ https://datatracker.ietf.org/doc/draft-pouwelse-perpass-shad... Key problem is that people are re-inventing the wheel in 100+ projects. Good scalable code takes year to develop. The mesh project I'm working on has taken 42 man-years to date :https://github.com/Tribler/tribler/wiki https://github.com/Tribler/tribler/wiki
- huwtheperson 13y ago
  Ugh, tell me about it. Every time one of these comes around I feel sorry for the poor guys at [CJDNS](https://github.com/cjdelisle/cjdns https://github.com/cjdelisle/cjdns). They've been working on it for ages now and the whitepaper behind it is amazing. Sadly, not enough attention, as I fear most of these mesh networks get.
  - synctext 13y ago
    mesh/overlay stuff has been developed for decades, we even have lists of projects:http://redecentralize.org/interviews/ http://redecentralize.org/interviews/ https://github.com/redecentralize/alternative-internet https://github.com/redecentralize/alternative-internet Why don't we work more together?
    - morphle 13y ago
      Very very good point. My research group and ISP federation tries to work together with all of these 100+ initiatives, but sadly, a large fraction suffer from the "not invented here" syndrome or even worse, the "never read any scientific papers that solved this many years ago" syndrome. (See my profile).
      - femto 13y ago
        I read your profile, and looked though the websites that it mentioned, but I'm not really any clearer on the connection to mesh networking. Do you mind giving a summary? I genuinely am interested. I'm interested in mesh networking and its relation to the physical layer. As a (time deficient) hobby, I've been noodling on a bit of free space optics for a number of years. In my day job, I design physical layer radio infrastructure for data/voice networks. (I just realised that "noodling" probably doesn't make sense outside Au/NZ. It comes from the practice of picking though mine tailings for opals, meaning: to toil in an ad hoc manner over a long period of time.)
tlrobinson 13y ago
For me, one of the most exciting things about Bitcoin is that it can be used to solve incentivisation problems in decentralized/p2p applications like this, without relying on a centralized 3rd party. BitTorrent has done well essentially using bandwidth as a currency, but it still relies on some users giving away more bandwidth than they consume, which doesn't necessarily work with other applications.
- abdullahkhalids 13y ago
  I will argue against this. Having a currency that can be bought or sold is also an incentive for big central service providers to crop up. For instance, bitcoin is not as decentralized as it should have been with only a few large mining pools.
- lvs 13y ago
  just to be clear, the "bitcoin" in this proposal is a token of network traffic, more like your example of bittorrent than bitcoin really, with the caveat that it's built using an architecture more like bitcoin's. the incentives are simply based on a mutual desire to exchange traffic, like bittorrent. there's no monetary incentive, and i shudder to think what the consequences would be of thinking about network traffic as a traded commodity on some kind of market or exchange. (oh god that's almost the status quo!)
lotsofmangos 13y ago
It should really say "proposed" somewhere in the title.
- imdsm 13y ago
  Totally agree. Sounds great, but I'd like to know more about who is behind it.
  - lotsofmangos 13y ago
    Don't get me wrong, I really like some of the concepts they are considering for the architecture, however until there is a working beta, I'm just going to file this as vapourware. I am always a little suspicious of publicising something as a solution before a prototype exists.
    - jnbiche 13y ago
      This is a problem with the Bitcoin community. I think enthusiastic people come up with a good idea and release a white paper, thinking others will jump on board to help make a prototype. They forget that Satoshi released a white paper, then put his nose to the grindstone for 2-3 months before releasing a working implementation.
      - kylebrown 13y ago
        According to Satoshi, he worked on the implementation before the whitepaper. iirc he released both at the same time.
        jnbiche 13y ago
        Well, he may have been working on it, but he released the Bitcoin implementation several months after he released the whitepaper. He released the white paper to the public on Oct 31/Nov. 1 2008, and the actual Windows client (the initial client) on Jan 3, 2009. I don't have time to dig up links at the moment, but you can check the cryptography list archives for confirmation.
michaelt 13y ago
I'm no expert on mesh networks so maybe someone can tell me - how would a mesh network connect Los Angeles and Las Vegas? Or San Francisco to Salt Lake City? Given the populations on both ends, I assume you'd need more bandwidth than you could get from wifi routers with pringles can antennas. I assume the companies who own long distance fiber have to do what the government says, as they need government permission to dig up the streets and install more fiber meaning they can't afford to upset politicians. So you can encrypt the traffic, but if the word comes down to cut the cable the cable gets cut.
- karl_gluck 13y ago
  Not too long ago, the internet measured top speeds in tens of kBps. So first off, give it time. Engineers are good at solving problems like this. Furthermore, mesh networking is a fundamentally different architecture solving a different problem than the internet. The current internet works great for high-bandwidth applications like commercial video streaming, but is terrible in terms of protecting itself from concerted attack by powerful actors. Mesh networks solve the single-point-of-failure problem. http://gizmodo.com/5912383/how-to-destroy-the-internet http://gizmodo.com/5912383/how-to-destroy-the-internet
- ris 13y ago
  A mesh network is not necessarily the same thing as a wireless mesh network.
  - lappa 13y ago
    But that is how this mesh is supposed to work according to point 3.
    - ris 13y ago
      It being a mesh protocol would supposedly mean that cut cables would get routed around fairly automatically. You'd have to cut all the cables as there'd always be a financial incentive for someone to carry that data to its destination. And there is more than just wifi to pick up the slack in an emergency. Free space optical links, satellite uplinks are all possible mediums. I'm not arguing that this would work well, but I think this is the idea behind it all.
MWil 13y ago
does "bitcoin-based" have anything to do with the priority of delivery over the network?
- cmpaul 13y ago
  The way I'm reading it, nodes pay their neighbors for successfully delivered packets using bitcoin-like credits mined by specialized nodes. I'd imagine the nodes with high balances would then be prioritized for delivery (both sending/receiving).
  - ZenoArrow 13y ago
    If that's really how it works then my gut instinct tells me it's wrong. We shouldn't rely on any form of currency exchange for negotiating the movement of web traffic, otherwise we risk messing with net neutrality. Bandwidth should be balanced by tracking load, not by currency. Cryptocurrencies could prove useful as a form of identity on the network, a secure ID token that isn't traded but does allow nodes to recognise each other, I see the merit in that.
    - cmpaul 13y ago
      Yes, right. Not currency, identity. The bitcoin approach is used to uniquely identify specific transactions for a node's traffic.
      - morphle 13y ago
        But can these people built a better Tor network/tunnel than the Tor/EFF people did? Doubtful. Libernet seems based on Linux, itself not secure. My group has solved both this problem, without using security flawed Linux/Unix (I will be happy to send you the scientific papers proving my point. See my HN profile).
gojomo 13y ago
The 1st time I heard the analogy of a "silk road" online, it wasn't for an underground marketplace, but an agoric packet-routing scheme, the "Digital Silk Road": http://www.virtualschool.edu/mon/Bionomics/Extropians/HardyTribbleSilkRoad.html http://www.virtualschool.edu/mon/Bionomics/Extropians/HardyT... Under that scheme, local microcredit along each hop would (theoretically) allow global routing without any global/monopolistic governance or currency.
wcoenen 13y ago
Ripple (as it was originally conceived by Ryan Fugger) seems to be a much better match for this idea than bitcoin. The connections between nodes could work as the local credit relationships of Ripple.
jokoon 13y ago
I'd be more interested in a software that hosts data in a mesh network fashion rather than talk about zombie apocalypse, SOPA/PIPA and other political craze. And why does this thing talk so much about hardware anyways ? Maybe the goal is to get funding to make that kind of project happen, but I don't think any engineer will like to get involved into a project than explicitely express political opinions and concerns. Let the politics deal with the NSA stories so it can get solved. If you really want to make a difference like gnutella, edonkey and bittorrent did, just be a programmer and get working. I think the only thing that matter is to make a software protocol which is interesting to use, not a meli melo of nerd-oriented awesomeness.
- higherpurpose 13y ago
  You mean like MaidSafe? http://maidsafe.net/ http://maidsafe.net/ As Lawrence Lessig said, "code is law". We've seen it before. Governments are usually too behind in understanding what's going on. Sometimes it helps, like with bittorent and TPB, while other times it hurts us, like with DRM being able to lock stuff that shouldn't actually be locked according to the law, but the DRM makers are basically making their own laws through code, or with NSA who spies on anyone they want no matter what laws exist, because they will just twist them to mean whatever they want in their "interpretations" of the law - and then keep those interpretations secret for "national security". I don't think it's "political" to say you want anonymity or privacy to exist on a network, or that you don't want a network to be censored. You say "just program like the Bittorrent programmers did". Isn't that what they're doing? Bittorrent appeared because Napster and Kazaa could eventually be censored and stopped. Now we need new networks and apps that are decentralized by default. These also bring increased security for everyone, increased privacy, and more resilience against DDOS attacks, too. So I'm not sure what you're really attacking here.
  - jokoon 13y ago
    > I don't think it's "political" to say you want anonymity or privacy to exist on a network anonymity is just a tool you use to protect your privacy if it's being prone to attacks. Pseudominity is fine most of the time. > Isn't that what they're doing? > So I'm not sure what you're really attacking here. Bittorrent or previous protocols were not advertising their project as some new idea that still needed to be implemented. Previously they just released their apps, people used them, the app evolved and it grew more popular. I think I see too many of those project embryos while there already is freenet and tor. It's either an attempt from investor to attracts tech unsavy people into a security trap, or inexperienced or tech unsavy project initiators. Maybe there is more work to be done to make freenet and tor more accessible to new users instead of reinventing new techs. Nowadays with the sharp increase of internet users, the noose is indeed getting tighter and tighter. It doesn't mean you have to reinvent the wheel... Improve it instead.
flibertgibit 13y ago
I love the idea of mesh networking to decentralize the net and using OTS solutions to do so. In fact, I've pushed for it on HN. But, here are a few problems: * Before a "theory" website with cheesy blog/helloworld/contact images http://openlibernet.org/index.html#contribute http://openlibernet.org/index.html#contribute, come up with a solution. The U.S. has been through several years of executive and legislative branches led on both sides by activists with good intentions and no real, thought-out workable solutions. Beyond bumper stickers of hope, you must have solutions, truly sell them, compromise on some keys points so you have enough people on board, and implement. We don't need a page telling us about a net we can't use yet. * Bitcoin. Don't tie something completely unrelated to a networking solution. * To make it work (really), you need points on the network that can do short and long-range. Wifi only is not enough.
- higherpurpose 13y ago
  * Bitcoin. Don't tie something completely unrelated to a networking solution. Why not? We've seen plenty of good ideas that were inspired by Bitcoin (P2P DNS system, platform with a built-in Turing complete language, true P2P Twitter alternative for the first time, encrypted e-mail without PGP, P2P back-up solutions where you pay others to keep your data, rather than on a centralized place that can be hacked or abused, etc), and we're going to see a lot more like this. The only problem is that unfortunately Bitcoin was never meant to be a platform on which to build other "apps". This is why all these "Bitcoin-based" or Bitcoin-inspired projects are actually heavily modified versions of Bitcoin, and not apps that work on top of Bitcoin. But I think Ethereum will solve this by creating another "Bitcoin-like" platform on which you can build anything you want, and everything is P2P by default. http://www.ethereum.org/ http://www.ethereum.org/
- troymc 13y ago
  They aren't using Bitcoin. They're using the Bitcoin Protocol but to trade in something else, which they call "Traffic." I suppose you'd be able to trade Traffic for Bitcoin, US dollars, or cabbage, once exchanges get set up.
- wmf 13y ago
  I don't think incentives are completely unrelated to networking. They may not have the right design, but in theory applying Bitcoin-like concepts to networking incentives isn't a crazy idea.
gus_massa 13y ago
I don't like that they use misleading information to exaggerate the problems of the current Internet: > One such limitation is the IPv4 protocol. With a 32 bit IP Address, it is only inevitable that the address pool will be exhausted (as it almost has) as the internet keeps on expanding. This is true, but they completely ignore the advance of IPv6.
- OpenLibernet 13y ago
  In order for IPv6 to ever become reality, a massive upgrade of infrastructure is required. So far ISPs have been reluctant to make the move. Whether or not ipv6 will ever see the light of day is arguable. We think not.
rsync 13y ago
I don't see this in the FAQ ... What if I create an entire libernet network (say, 8 nodes, all meshed) and isolate, or fork them, from the rest of libernet ... and then generate massive traffic between them all to "earn" traffic currency ? So, 8 nodes all connected with gigabit ethernet, all maxing out their traffic in one big circle, earning "traffic" currency ... and then I rejoin the mainline libernet in the future with my massive store of traffic currency. How do you avoid that scenario ? Certainly you can't make the network inoperable (or non-earning) if a network link breaks somewhere in the mesh ... because there would always be links in the mesh breaking. Is there a minimum number of peers accessible in order to earn traffic currency ? Is that number larger than the number of freebsd jails I could quickly initialize with a script ?
- wyager 13y ago
  This is why the only two options we have right now are a) volunteer-based carriers and b) market-based carriers. Some in-between probably won't work very well.
- saurik 13y ago
  One would imagine that the whole point of working with a protocol similar to bitcoin is that this is not possible: you can't fork the network and rejoin at a later time; that would be modeled as a blockchain fork, and your chain would be shorter than the main chain when you tried to rejoin the rest of the peers and your history would be erased.
  - rsync 13y ago
    Sure, but that doesn't make any sense in the context of a network. You're saying that if my link goes down, or if my neighborhoods link goes own ... or if my city link goes down ... I can never rejoin the network ? Or perhaps I can never spend my traffic currency ? Of course the links will go down all the time and various (large and small) segments of the network will be "islands" from time to time (as parts of the Internet sometimes are) ... and I wonder how currency accumulation behaves at these times.
    - 13y ago
jccooper 13y ago
A proof-of-stake system (see PeerCoin) might be nicer than encouraging Bitcoin-style proof-of-work mining throughout the network. Possibly it could even be run on simple nodes, as it doesn't require gobs of processing power.
- goodside 13y ago
  Bitcoin only requires extreme processing power if you're trying mine blocks. For a node that just wants to issue and monitor transactions, there's no resource advantage to using Peercoin or other proof-of-stake coins.
OpenLibernet 13y ago
To clear some things up: 1) The theory itself is a WIP. We only published in the hopes of getting reviews, comments and ideas from the community 2) The last thing we want to do is reinvent the wheel 3) We do not bring any novelty to either mesh routing or p2p payment systems. We only combined the two to create a mesh protocol with a built-in monetary incentive 4) The actual purpose of the monetary incentive is to get corporations capable of installing fiber links interested; we do understand it is impossible to create a global wireless mesh network. 5) Tunnel nodes are needed until a stable (fiber) infrastructure is built to replace them 6) It is always possible for other mesh network projects to implement the payment system into their protocols; we did not discover fire, we only proposed a solution to what we believe to be the main obstacle to widespread adoption of mesh networks; i.e. lack of incentive. 7) There are no political motivations for the project. Demanding a secure and spying-free communication channel is not anarchism. It is only the natural evolution of the internet. AFAIK eavesdropping is not just confined to the government/nsa. Even small ISPs possess the ability to spy on their clients. Also note that there are countries outside the US that have completely different concerns in regards to internet laws and regulations. 8) We understand there may be serious flaws in the theory at it's current state which is why we're teaming up with industry experts to refine and improve it.
- ofxartem 13y ago
  Meshnet has been doing this for some time: https://projectmeshnet.org/ https://projectmeshnet.org/. There are also other groups of people fighting for the same thing in Germany: http://en.wikipedia.org/wiki/Freifunk http://en.wikipedia.org/wiki/Freifunk
  - OpenLibernet 13y ago
    The Batman protocol doesn't scale well. Meshnet doesn't have the incentive (the actual only thing we're contributing to mesh networking) If we could apply our payment strategy to any of the existing protocols we would have done it. Also, the fact that we have incentive means we could add special nodes; which means we can solve some of the issues that other protocols suffer from more easily. Instead of relying on distributed hash tables for example (a layer of complexity), why not dedicate a special node for lookup queries (PLN). That and we don't mind adopting a completely different approach for the routing protocol. This is why we're asking for the help of the community. Please read the paper.
pistle 13y ago
Infrastructure is hand-waving friendly? The revolution will not be Pringle-can-ized. Is this how bad it's getting for bitcoin? People are trying to invent dependencies based on flying currencies? Let me get this straight... You want me to leave my darknet router on and running to help route all that traffic from all that blockchain crap related to bitcoin transactions? And my incentive is to get infinitesimally small amounts of this stuff (which then requires more blockchain network overhead)??? What is the likely ratio of blockchain noise to real data signal? I enter the coffee shop and then we bucket brigade my order to the counter. Anyone standing between me and the barista will have to shake hands, exchange pleasantries, trade 1 hundred-millionth of a bitcoin, check some blockchains, express their love of the community that makes the ordering of coffee possible, then pass my order through letter-by-letter. Once complete, the same process occurs in reverse, except the coffee will now be passed drop by drop from mouth to mouth of everybody between the barista and myself (swished around to ensure proper encryption). Let's not also forget that I'm bootstrapping a fiber optic network through my neighbors' back yards, because if I stick my router behind my Comcast modem... uh, yeah. Wifi? pshaw. My neighbors can't change a default password on their router. Now they're my decentralized network engineers that "love the community"?? Good lord, the bitbug miners are trying to take their electricity wasting exercise and make it a network wasting exercise as well.

126 points by gasull 13y ago

A large scale mesh network is interesting to me for a couple of reasons: * Consumers would no longer be dependent on monopolies like Comcast for internet service (OSI layer 1) * It is much more resistant to eavesdropping and censorship (OSI layers 2-3) These seem like two totally different problems to me, and I'm still struggling to understand how the physical layer part can work on a large scale. I get how they could work on a small scale. For example, you could easily set up an ad-hoc wifi network in a college dorm for, say, file sharing purposes. But in most other settings, I don't get how it could work. Most people don't even talk to their neighbors, so even if you lived in a neighborhood that had a high enough density of nodes for this to work, why would you want to share a little ad-hoc network with your neighbors? Of course, wireless networks alone aren't going to be sufficient to build a large-scale mesh network that could disrupt existing ISPs. The latency alone makes this infeasible. So you need someone to invest in physical links, especially physical links that cover long distances. That's why the BTC part of this proposal is so great -- it provides a way to compensate these people. Of course, this is just a slightly different problem than the one we have today, but instead of being stuck with Comcast we're stuck with whoever runs these physical links.

lotsofmangos 13y ago
wireless networks alone aren't going to be sufficient to build a large-scale mesh network that could disrupt existing ISPs. The latency alone makes this infeasible. I'm not so sure, your links are almost c, whereas optic fiber is substantially slower, so it depends entirely on how little processing you can get away with in the message passing code. edit - was just thinking about this, an arm cortex runs at ~ 2,000 MIPS, which gives you just under 15 cm per instruction for radio in air. Fiber is about a third slower, so your links do not actually have to be massively long to give you quite a few cycles to play with.
- jsaxton86 13y ago
  I think it mostly depends upon the range of each of the nodes. If your nodes have an average range of 820 feet (the range of 802.11n (in good weather, I assume)), and you have 1 node exactly every 820 feet between NY and SF, that's over 16,000 links. If each link has 10ms of latency, that's 16s of latency.
  - ZenoArrow 13y ago
    16s of latency is interesting to think about, do think we'll get around this issue with local web page caching (obviously won't work for non turn-based games). SyncNet or something similar could prove very useful... http://jack.minardi.org/software/syncnet-a-decentralized-web-browser/ http://jack.minardi.org/software/syncnet-a-decentralized-web...
- ris 13y ago
  It's got nothing to do with the propagation speed. Wireless has high latency because it requires comparatively high amounts of processing, has to operate over a shared channel (requiring collision avoidance), is subject to unexpected interference...
  - adestefan 13y ago
    You also need to send a lot of additional information for error correction.
JoshTriplett 13y ago
> It is much more resistant to eavesdropping and censorship (OSI layers 2-3) Not inherently. Arguably, a large-scale mesh network would be more susceptible to interference, due to the larger number of distinct parties involved in each communication. The difference is that because a mesh network assumes the lack of trust in the intermediate nodes, a sensible design for one has to include end-to-end security.
rmrfrmrf 13y ago
I think you have your OSI layers mixed up.
jewel 13y ago
> Most people don't even talk to their neighbors, so even if you lived in a neighborhood that had a high enough density of nodes for this to work, why would you want to share a little ad-hoc network with your neighbors? I've been thinking about this recently. The idea is that it's transparent. Imagine, for example, if Apple started shipping this on all iOS devices, or it became a common android feature. Here is a blog post I wrote on the topic last month: http://stevenjewel.com/2014/01/android-mesh/ http://stevenjewel.com/2014/01/android-mesh/

tinco 13y ago

Has no reference implementation.

cmpaul 13y ago
"20. What is the status of Open Libernet at this time? We are still at idea stage. We’ve built a few simulations to test some parts of the system, but the main development hasn’t begun. We have published our whitepaper, which explains – to some degree of detail – the proposed protocol, and we’re still waiting for review. For that, if you happen to have any insight on routing protocols, networking or software development that you would like to share with us, please do so. If you think something with the protocol is not right or needs to be improved, or if you have a better idea on how to do something, also do not hesitate to voice your opinion. This is a community project and we are more than happy to receive all the help we could get. Also, everything we’re doing is Open Source, including the white-paper, the concept and in the future the code. You are free to take anything you like, modify, publish and release, then tell us about it. Or if you prefer, don’t."

deanclatworthy 13y ago

I read these mesh network posts with great interest. I really like the idea of a neighbourhood having their own private network for communication. Sadly this project is not yet ready. Let's say I found 10 people in my neighbourhood with 10 routers what's out there that's easy to set up that would get us started with a little mesh network?

synctext 13y ago

This is already moving within IETF: https://datatracker.ietf.org/doc/draft-pouwelse-perpass-shadow-internet/ https://datatracker.ietf.org/doc/draft-pouwelse-perpass-shad... Key problem is that people are re-inventing the wheel in 100+ projects. Good scalable code takes year to develop. The mesh project I'm working on has taken 42 man-years to date :https://github.com/Tribler/tribler/wiki https://github.com/Tribler/tribler/wiki

huwtheperson 13y ago
Ugh, tell me about it. Every time one of these comes around I feel sorry for the poor guys at [CJDNS](https://github.com/cjdelisle/cjdns https://github.com/cjdelisle/cjdns). They've been working on it for ages now and the whitepaper behind it is amazing. Sadly, not enough attention, as I fear most of these mesh networks get.
- synctext 13y ago
  mesh/overlay stuff has been developed for decades, we even have lists of projects:http://redecentralize.org/interviews/ http://redecentralize.org/interviews/ https://github.com/redecentralize/alternative-internet https://github.com/redecentralize/alternative-internet Why don't we work more together?
  - morphle 13y ago
    Very very good point. My research group and ISP federation tries to work together with all of these 100+ initiatives, but sadly, a large fraction suffer from the "not invented here" syndrome or even worse, the "never read any scientific papers that solved this many years ago" syndrome. (See my profile).
    - femto 13y ago
      I read your profile, and looked though the websites that it mentioned, but I'm not really any clearer on the connection to mesh networking. Do you mind giving a summary? I genuinely am interested. I'm interested in mesh networking and its relation to the physical layer. As a (time deficient) hobby, I've been noodling on a bit of free space optics for a number of years. In my day job, I design physical layer radio infrastructure for data/voice networks. (I just realised that "noodling" probably doesn't make sense outside Au/NZ. It comes from the practice of picking though mine tailings for opals, meaning: to toil in an ad hoc manner over a long period of time.)

tlrobinson 13y ago

For me, one of the most exciting things about Bitcoin is that it can be used to solve incentivisation problems in decentralized/p2p applications like this, without relying on a centralized 3rd party. BitTorrent has done well essentially using bandwidth as a currency, but it still relies on some users giving away more bandwidth than they consume, which doesn't necessarily work with other applications.

abdullahkhalids 13y ago
I will argue against this. Having a currency that can be bought or sold is also an incentive for big central service providers to crop up. For instance, bitcoin is not as decentralized as it should have been with only a few large mining pools.
lvs 13y ago
just to be clear, the "bitcoin" in this proposal is a token of network traffic, more like your example of bittorrent than bitcoin really, with the caveat that it's built using an architecture more like bitcoin's. the incentives are simply based on a mutual desire to exchange traffic, like bittorrent. there's no monetary incentive, and i shudder to think what the consequences would be of thinking about network traffic as a traded commodity on some kind of market or exchange. (oh god that's almost the status quo!)

lotsofmangos 13y ago

It should really say "proposed" somewhere in the title.

imdsm 13y ago
Totally agree. Sounds great, but I'd like to know more about who is behind it.
- lotsofmangos 13y ago
  Don't get me wrong, I really like some of the concepts they are considering for the architecture, however until there is a working beta, I'm just going to file this as vapourware. I am always a little suspicious of publicising something as a solution before a prototype exists.
  - jnbiche 13y ago
    This is a problem with the Bitcoin community. I think enthusiastic people come up with a good idea and release a white paper, thinking others will jump on board to help make a prototype. They forget that Satoshi released a white paper, then put his nose to the grindstone for 2-3 months before releasing a working implementation.
    - kylebrown 13y ago
      According to Satoshi, he worked on the implementation before the whitepaper. iirc he released both at the same time.
      - jnbiche 13y ago
        Well, he may have been working on it, but he released the Bitcoin implementation several months after he released the whitepaper. He released the white paper to the public on Oct 31/Nov. 1 2008, and the actual Windows client (the initial client) on Jan 3, 2009. I don't have time to dig up links at the moment, but you can check the cryptography list archives for confirmation.

michaelt 13y ago

I'm no expert on mesh networks so maybe someone can tell me - how would a mesh network connect Los Angeles and Las Vegas? Or San Francisco to Salt Lake City? Given the populations on both ends, I assume you'd need more bandwidth than you could get from wifi routers with pringles can antennas. I assume the companies who own long distance fiber have to do what the government says, as they need government permission to dig up the streets and install more fiber meaning they can't afford to upset politicians. So you can encrypt the traffic, but if the word comes down to cut the cable the cable gets cut.

karl_gluck 13y ago
Not too long ago, the internet measured top speeds in tens of kBps. So first off, give it time. Engineers are good at solving problems like this. Furthermore, mesh networking is a fundamentally different architecture solving a different problem than the internet. The current internet works great for high-bandwidth applications like commercial video streaming, but is terrible in terms of protecting itself from concerted attack by powerful actors. Mesh networks solve the single-point-of-failure problem. http://gizmodo.com/5912383/how-to-destroy-the-internet http://gizmodo.com/5912383/how-to-destroy-the-internet
ris 13y ago
A mesh network is not necessarily the same thing as a wireless mesh network.
- lappa 13y ago
  But that is how this mesh is supposed to work according to point 3.
  - ris 13y ago
    It being a mesh protocol would supposedly mean that cut cables would get routed around fairly automatically. You'd have to cut all the cables as there'd always be a financial incentive for someone to carry that data to its destination. And there is more than just wifi to pick up the slack in an emergency. Free space optical links, satellite uplinks are all possible mediums. I'm not arguing that this would work well, but I think this is the idea behind it all.

MWil 13y ago

does "bitcoin-based" have anything to do with the priority of delivery over the network?

cmpaul 13y ago
The way I'm reading it, nodes pay their neighbors for successfully delivered packets using bitcoin-like credits mined by specialized nodes. I'd imagine the nodes with high balances would then be prioritized for delivery (both sending/receiving).
- ZenoArrow 13y ago
  If that's really how it works then my gut instinct tells me it's wrong. We shouldn't rely on any form of currency exchange for negotiating the movement of web traffic, otherwise we risk messing with net neutrality. Bandwidth should be balanced by tracking load, not by currency. Cryptocurrencies could prove useful as a form of identity on the network, a secure ID token that isn't traded but does allow nodes to recognise each other, I see the merit in that.
  - cmpaul 13y ago
    Yes, right. Not currency, identity. The bitcoin approach is used to uniquely identify specific transactions for a node's traffic.
    - morphle 13y ago
      But can these people built a better Tor network/tunnel than the Tor/EFF people did? Doubtful. Libernet seems based on Linux, itself not secure. My group has solved both this problem, without using security flawed Linux/Unix (I will be happy to send you the scientific papers proving my point. See my HN profile).

gojomo 13y ago

The 1st time I heard the analogy of a "silk road" online, it wasn't for an underground marketplace, but an agoric packet-routing scheme, the "Digital Silk Road": http://www.virtualschool.edu/mon/Bionomics/Extropians/HardyTribbleSilkRoad.html http://www.virtualschool.edu/mon/Bionomics/Extropians/HardyT... Under that scheme, local microcredit along each hop would (theoretically) allow global routing without any global/monopolistic governance or currency.

wcoenen 13y ago

Ripple (as it was originally conceived by Ryan Fugger) seems to be a much better match for this idea than bitcoin. The connections between nodes could work as the local credit relationships of Ripple.

jokoon 13y ago

I'd be more interested in a software that hosts data in a mesh network fashion rather than talk about zombie apocalypse, SOPA/PIPA and other political craze. And why does this thing talk so much about hardware anyways ? Maybe the goal is to get funding to make that kind of project happen, but I don't think any engineer will like to get involved into a project than explicitely express political opinions and concerns. Let the politics deal with the NSA stories so it can get solved. If you really want to make a difference like gnutella, edonkey and bittorrent did, just be a programmer and get working. I think the only thing that matter is to make a software protocol which is interesting to use, not a meli melo of nerd-oriented awesomeness.

higherpurpose 13y ago
You mean like MaidSafe? http://maidsafe.net/ http://maidsafe.net/ As Lawrence Lessig said, "code is law". We've seen it before. Governments are usually too behind in understanding what's going on. Sometimes it helps, like with bittorent and TPB, while other times it hurts us, like with DRM being able to lock stuff that shouldn't actually be locked according to the law, but the DRM makers are basically making their own laws through code, or with NSA who spies on anyone they want no matter what laws exist, because they will just twist them to mean whatever they want in their "interpretations" of the law - and then keep those interpretations secret for "national security". I don't think it's "political" to say you want anonymity or privacy to exist on a network, or that you don't want a network to be censored. You say "just program like the Bittorrent programmers did". Isn't that what they're doing? Bittorrent appeared because Napster and Kazaa could eventually be censored and stopped. Now we need new networks and apps that are decentralized by default. These also bring increased security for everyone, increased privacy, and more resilience against DDOS attacks, too. So I'm not sure what you're really attacking here.
- jokoon 13y ago
  > I don't think it's "political" to say you want anonymity or privacy to exist on a network anonymity is just a tool you use to protect your privacy if it's being prone to attacks. Pseudominity is fine most of the time. > Isn't that what they're doing? > So I'm not sure what you're really attacking here. Bittorrent or previous protocols were not advertising their project as some new idea that still needed to be implemented. Previously they just released their apps, people used them, the app evolved and it grew more popular. I think I see too many of those project embryos while there already is freenet and tor. It's either an attempt from investor to attracts tech unsavy people into a security trap, or inexperienced or tech unsavy project initiators. Maybe there is more work to be done to make freenet and tor more accessible to new users instead of reinventing new techs. Nowadays with the sharp increase of internet users, the noose is indeed getting tighter and tighter. It doesn't mean you have to reinvent the wheel... Improve it instead.

flibertgibit 13y ago

I love the idea of mesh networking to decentralize the net and using OTS solutions to do so. In fact, I've pushed for it on HN. But, here are a few problems: * Before a "theory" website with cheesy blog/helloworld/contact images http://openlibernet.org/index.html#contribute http://openlibernet.org/index.html#contribute, come up with a solution. The U.S. has been through several years of executive and legislative branches led on both sides by activists with good intentions and no real, thought-out workable solutions. Beyond bumper stickers of hope, you must have solutions, truly sell them, compromise on some keys points so you have enough people on board, and implement. We don't need a page telling us about a net we can't use yet. * Bitcoin. Don't tie something completely unrelated to a networking solution. * To make it work (really), you need points on the network that can do short and long-range. Wifi only is not enough.

higherpurpose 13y ago
* Bitcoin. Don't tie something completely unrelated to a networking solution. Why not? We've seen plenty of good ideas that were inspired by Bitcoin (P2P DNS system, platform with a built-in Turing complete language, true P2P Twitter alternative for the first time, encrypted e-mail without PGP, P2P back-up solutions where you pay others to keep your data, rather than on a centralized place that can be hacked or abused, etc), and we're going to see a lot more like this. The only problem is that unfortunately Bitcoin was never meant to be a platform on which to build other "apps". This is why all these "Bitcoin-based" or Bitcoin-inspired projects are actually heavily modified versions of Bitcoin, and not apps that work on top of Bitcoin. But I think Ethereum will solve this by creating another "Bitcoin-like" platform on which you can build anything you want, and everything is P2P by default. http://www.ethereum.org/ http://www.ethereum.org/
troymc 13y ago
They aren't using Bitcoin. They're using the Bitcoin Protocol but to trade in something else, which they call "Traffic." I suppose you'd be able to trade Traffic for Bitcoin, US dollars, or cabbage, once exchanges get set up.
wmf 13y ago
I don't think incentives are completely unrelated to networking. They may not have the right design, but in theory applying Bitcoin-like concepts to networking incentives isn't a crazy idea.

gus_massa 13y ago

I don't like that they use misleading information to exaggerate the problems of the current Internet: > One such limitation is the IPv4 protocol. With a 32 bit IP Address, it is only inevitable that the address pool will be exhausted (as it almost has) as the internet keeps on expanding. This is true, but they completely ignore the advance of IPv6.

OpenLibernet 13y ago
In order for IPv6 to ever become reality, a massive upgrade of infrastructure is required. So far ISPs have been reluctant to make the move. Whether or not ipv6 will ever see the light of day is arguable. We think not.

rsync 13y ago

I don't see this in the FAQ ... What if I create an entire libernet network (say, 8 nodes, all meshed) and isolate, or fork them, from the rest of libernet ... and then generate massive traffic between them all to "earn" traffic currency ? So, 8 nodes all connected with gigabit ethernet, all maxing out their traffic in one big circle, earning "traffic" currency ... and then I rejoin the mainline libernet in the future with my massive store of traffic currency. How do you avoid that scenario ? Certainly you can't make the network inoperable (or non-earning) if a network link breaks somewhere in the mesh ... because there would always be links in the mesh breaking. Is there a minimum number of peers accessible in order to earn traffic currency ? Is that number larger than the number of freebsd jails I could quickly initialize with a script ?

wyager 13y ago
This is why the only two options we have right now are a) volunteer-based carriers and b) market-based carriers. Some in-between probably won't work very well.
saurik 13y ago
One would imagine that the whole point of working with a protocol similar to bitcoin is that this is not possible: you can't fork the network and rejoin at a later time; that would be modeled as a blockchain fork, and your chain would be shorter than the main chain when you tried to rejoin the rest of the peers and your history would be erased.
- rsync 13y ago
  Sure, but that doesn't make any sense in the context of a network. You're saying that if my link goes down, or if my neighborhoods link goes own ... or if my city link goes down ... I can never rejoin the network ? Or perhaps I can never spend my traffic currency ? Of course the links will go down all the time and various (large and small) segments of the network will be "islands" from time to time (as parts of the Internet sometimes are) ... and I wonder how currency accumulation behaves at these times.
  - 13y ago

jccooper 13y ago

A proof-of-stake system (see PeerCoin) might be nicer than encouraging Bitcoin-style proof-of-work mining throughout the network. Possibly it could even be run on simple nodes, as it doesn't require gobs of processing power.

goodside 13y ago
Bitcoin only requires extreme processing power if you're trying mine blocks. For a node that just wants to issue and monitor transactions, there's no resource advantage to using Peercoin or other proof-of-stake coins.

OpenLibernet 13y ago

To clear some things up: 1) The theory itself is a WIP. We only published in the hopes of getting reviews, comments and ideas from the community 2) The last thing we want to do is reinvent the wheel 3) We do not bring any novelty to either mesh routing or p2p payment systems. We only combined the two to create a mesh protocol with a built-in monetary incentive 4) The actual purpose of the monetary incentive is to get corporations capable of installing fiber links interested; we do understand it is impossible to create a global wireless mesh network. 5) Tunnel nodes are needed until a stable (fiber) infrastructure is built to replace them 6) It is always possible for other mesh network projects to implement the payment system into their protocols; we did not discover fire, we only proposed a solution to what we believe to be the main obstacle to widespread adoption of mesh networks; i.e. lack of incentive. 7) There are no political motivations for the project. Demanding a secure and spying-free communication channel is not anarchism. It is only the natural evolution of the internet. AFAIK eavesdropping is not just confined to the government/nsa. Even small ISPs possess the ability to spy on their clients. Also note that there are countries outside the US that have completely different concerns in regards to internet laws and regulations. 8) We understand there may be serious flaws in the theory at it's current state which is why we're teaming up with industry experts to refine and improve it.

ofxartem 13y ago
Meshnet has been doing this for some time: https://projectmeshnet.org/ https://projectmeshnet.org/. There are also other groups of people fighting for the same thing in Germany: http://en.wikipedia.org/wiki/Freifunk http://en.wikipedia.org/wiki/Freifunk
- OpenLibernet 13y ago
  The Batman protocol doesn't scale well. Meshnet doesn't have the incentive (the actual only thing we're contributing to mesh networking) If we could apply our payment strategy to any of the existing protocols we would have done it. Also, the fact that we have incentive means we could add special nodes; which means we can solve some of the issues that other protocols suffer from more easily. Instead of relying on distributed hash tables for example (a layer of complexity), why not dedicate a special node for lookup queries (PLN). That and we don't mind adopting a completely different approach for the routing protocol. This is why we're asking for the help of the community. Please read the paper.

pistle 13y ago

Infrastructure is hand-waving friendly? The revolution will not be Pringle-can-ized. Is this how bad it's getting for bitcoin? People are trying to invent dependencies based on flying currencies? Let me get this straight... You want me to leave my darknet router on and running to help route all that traffic from all that blockchain crap related to bitcoin transactions? And my incentive is to get infinitesimally small amounts of this stuff (which then requires more blockchain network overhead)??? What is the likely ratio of blockchain noise to real data signal? I enter the coffee shop and then we bucket brigade my order to the counter. Anyone standing between me and the barista will have to shake hands, exchange pleasantries, trade 1 hundred-millionth of a bitcoin, check some blockchains, express their love of the community that makes the ordering of coffee possible, then pass my order through letter-by-letter. Once complete, the same process occurs in reverse, except the coffee will now be passed drop by drop from mouth to mouth of everybody between the barista and myself (swished around to ensure proper encryption). Let's not also forget that I'm bootstrapping a fiber optic network through my neighbors' back yards, because if I stick my router behind my Comcast modem... uh, yeah. Wifi? pshaw. My neighbors can't change a default password on their router. Now they're my decentralized network engineers that "love the community"?? Good lord, the bitbug miners are trying to take their electricity wasting exercise and make it a network wasting exercise as well.