10 ms·

Filecoin – Data storage network and crypto-currency based on Bitcoin

281 points by _prometheus 12y ago

wyager 12y ago
Is this one also vulnerable to people creating fake nodes that "store files" for each other to earn coins for doing nothing?
- dutchbrit 12y ago
  Exactly my thought! Seems like a big flaw.
- bdcs 12y ago
  The authors address Sybil attacks in the paper. Essentially, there is a PoW style consensus using hashes of files+metadata (as opposed to bitcoin which is hashes of nothing + metadata)
  - abc123xyz 12y ago
    whats to stop the node replying with the hash of the file saying "yes the file is here" when in reality no file is there
    - ZitchDog 12y ago
      Haven't read the paper, but could you ask for a hash of a random offset+byte count of the file?
      - slipstream- 12y ago
        Exactly what I was thinking.
      - wyager 12y ago
        That only works on direct person-to-person systems, not decentralized systems where the entire network has to reach consensus.
    - _prometheus 12y ago
      the proof-of-retrievability. see http://cseweb.ucsd.edu/~hovav/dist/verstore.pdf http://cseweb.ucsd.edu/~hovav/dist/verstore.pdf
      - philsnow 12y ago
        I'm at work and don't want to read either of the papers at the moment, but if A asks B to provide proof of having stored some bytes, could B proxy that request to C and replay the reply to A ?
        justin_ 12y ago
        As long as a node you control has the data and can retrieve the file, I don't think there's a problem. If you could proxy the "proof of retrieval" than you could surely proxy the retrieval itself. Pooling resources shouldn't break anything.
        sp332 12y ago
        But you could make it look like several computers have copies of the data when there's really only one copy of the data. You would get credit/coins for multiple copies and reduce the redundancy of the files.
        anonymousDan 12y ago
        I haven't RTFA or any of the papers but couldn't you perhaps add a different nonce to each copy of the file you want stored before encrypting?
        justin_ 12y ago
        From what I read in the paper, it sounds like you _don't_ get paid for providing the data. You get paid for proving you have it when mining a block. Distributing your storage across multiples nodes would simply make it easier to "mine", but wouldn't get coins passively. That is, if I'm understanding the protocol correctly.
- _prometheus 12y ago
  No, files are contributed to the network as a whole, must be distributed, and require payment. See the discussions on mining and on rewards in the paper: http://filecoin.io/filecoin.pdf http://filecoin.io/filecoin.pdf
  - JVIDEL 12y ago
    Hey sorry for the OT but shouldn't this be on the new Show section? just saying...
    - _prometheus 12y ago
      Well, it's not something you can TRY yet -- i think those are the guidelines? :)
flixic 12y ago
How to create new crypto-currency: - Pick something to base you currency on. Bit-, Name-, and File- are taken, but some might be left. - First letter, two stripes, here's a logo! - Publish a whitepaper. You must use LaTeX to typeset it. - No names or credit anywhere. Stay anonymous. Joking aside, this looks like something I will try. Have a spare 1TB connected to fast internet, and assume many others have as well.
- samirmenon 12y ago
  The Whois lookup of filecoin.io gives the name "Juan Batiz-Benet", so this isn't very anonymous.
  - _prometheus 12y ago
    nope, it isn't ;)
    - tinco 12y ago
      So why isn't your name on the paper? (not judging, just curious)
- _prometheus 12y ago
  Haha! Well, the crypto and game theory took a very long time to get right. (and more of it to come!) Send feedback!
jothi 12y ago
How different is it from Storjcoin ?
- mappum 12y ago
  Storj doesn't make any technical innovations, their plan is essentially "make some nice looking web apps and let other people worry about inventing the things we need". Storjcoin is a standard blockchain, but actually isn't even neccessary at all. Filecoin actually addresses the technical challenges in building something like this, and maybe Storj will end up using it in their stack.
  - ard10e 12y ago
    I spoke with super3, the Storj founder earlier today, and he was going over how they'll go about providing Proof of Storage. I'm not qualified to explain it myself, but expect a revised whitepaper in the next couple weeks that goes over it in detail.
- _prometheus 12y ago
  At first glance the concepts may seem similar, but this is a different thing altogether. Storj is a service layered on top of other blockchains. Filecoin is a new blockchain protocol itself, with its own contributions: mining construction, use of proof-of-retrievability, put/get as transactions, and more. Read: http://filecoin.io/filecoin.pdf http://filecoin.io/filecoin.pdf Soon will have simpler explanation. :)
- abc123xyz 12y ago
  Both smell like a scam, "coincidentally" today storj started fundraising by getting people to send them bitcoins for useless storjcoins they created, the worst part is people falling for such a scam, then again there's a fool born every minute which is a business model a lot of companies rely on. In this case and in storjcoin case the actors behind it all remain anonymous
  - _prometheus 12y ago
    I can understand why -- with all the scams out there -- that would be one reaction. But note that Filecoin makes important technical contributions on top of the Bitcoin protocol. You can read the paper here: http://filecoin.io/filecoin.pdf http://filecoin.io/filecoin.pdf FWIW, we aren't really anonymous. Neither in the storj case (you can find who they are in their website), nor in the Filecoin case. (hi!)
possibilistic 12y ago
Unrelated to Filecoin, this brought to mind a question about the Bitcoin blockchain I've been pondering recently. Since users can put small arbitrary byte data in a transaction that gets included in the block chain (AFAIK), isn't it possible that an attacker could insert illicit / illegal data that everyone using Bitcoin replicates? Copyrighted materials, sensitive personal information such as SSNs, and moreover, child pornograpy immediately come to mind as things that could forever be persisted. If this is possible, wouldn't that make everyone using the bitcoin blockchain criminally liable for storing and/or transmitting that information? (In the US and similar jurisdictions.) I assume lightweight clients that only store a hash would be unaffected. But still, this seems like a serious issue.
- wyager 12y ago
  This has already been done many times.
- _prometheus 12y ago
  Yes, it's been known for a while that people can upload illegal bits into blockchains. Don't think transaction steganography would hold up in court though. You could do similar things with DNS records, etc. File storage services though have the added concern that they're specifically designed to distribute files. Filecoin addresses this by (a) encouraging everyone to pre-encrypt before adding, and (b) creating a fluid market based on incentives. Nodes aren't required to store every file, only strongly incentivized to do so. Nodes are free to follow strong counter incentives (say, a well-known list of "illegal hashes" published by authorities). In reality, the way many courts handle things like this is to target the entry points: the websites advertising which are the illegal hashes and how to get them. (e.g. the piratebay instead of bittorrent inc.)
  - kang 12y ago
    Make encryption at source compulsory (since the data is uploaded through your software) and no need to tell the user about it. rather advertise it as a feature.
dubcanada 12y ago
Does "Renting disk space" not sound extremely sketchy to anyone else?
- _prometheus 12y ago
  Does "Renting your spare room" (airbnb) or "Renting your car" (uber) sound sketchy? It's all about how the protocol works and the security guarantees on top. Users should store pre-encrypted files. (except for public goods, like wikipedia)
Hilyin 12y ago
What about being liable for storing illegal files someone uploads to your filebox?
- elux 12y ago
  49310 83597 02850 19002 75777 67239 07649 57284 90777 21502 08632 08075 01840 97926 27885 09765 88645 57802 01366 00732 86795 44734 11283 17353 67831 20155 75359 81978 54505 48115 71939 34587 73300 38009 93261 95058 76452 50238 20408 11018 98850 42615 17657 99417 04250 88903 70291 19015 87003 04794 32826 07382 14695 41570 33022 79875 57681 89560 16240 30064 11151 69008 72879 83819 42582 71674 56477 48166 84347 92846 45809 29131 53186 00700 10043 35318 93631 93439 12948 60445 03709 91980 04770 94629 21558 18071 11691 53031 87628 84778 78354 15759 32891 09329 54473 50881 88246 54950 60005 01900 62747 05305 38116 42782 94267 47485 34965 25745 36815 11706 55028 19055 52656 22135 31463 10421 00866 28679 71144 46706 36692 19825 86158 11125 15556 50481 34207 68673 23407 65505 48591 08269 56266 69306 62367 99702 10481 23965 62518 00681 83236 53959 34839 56753 57557 53246 19023 48106 47009 87753 02795 61868 92925 38069 33052 04238 14996 99454 56945 77413 83356 89906 00587 08321 81270 48611 33682 02651 59051 66351 87402 90181 97693 93767 78529 28722 10955 04129 25792 57381 86605 84501 50552 50274 99477 18831 29310 45769 80909 15304 61335 94190 30258 81320 59322 77444 38525 50466 77902 45186 97062 62778 88919 79580 42306 57506 15669 83469 56177 97879 65920 16440 51939 96071 69811 12615 19561 02762 83233 98257 91423 32172 69614 43744 38105 64855 29348 87634 92103 09887 02878 74532 33132 53212 26786 33283 70279 25099 74996 94887 75936 91591 76445 88032 71838 47402 35933 02037 48885 06755 70658 79194 61134 19323 07814 85443 64543 75113 20709 86063 90746 41756 41216 35042 38800 29678 08558 67037 03875 09410 76982 11837 65499 20520 43682 55854 64228 85024 29963 32268 53691 24648 55000 75591 66402 47292 40716 45072 53196 74499 95294 48434 74190 21077 29606 82055 81309 23626 83798 79519 66199 79828 55258 87161 09613 65617 80745 66159 24886 60889 81645 68541 72136 29208 46656 27913 14784 66791 55096 51543 10113 53858 62081 96875 83688 35955 77893 91454 53935 68199 60988 08540 47659 07358 97289 89834 25047 12891 84162 65878 96821 85380 87956 27903 99786 29449 39760 54675 34821 25675 01215 17082 73710 76462 70712 46753 21024 83678 15940 00875 05452 54353 7. ^ https://en.wikipedia.org/wiki/Illegal_prime https://en.wikipedia.org/wiki/Illegal_prime Now what?
  - TomGullen 12y ago
slg 12y ago
This is pretty interesting from a economic standpoint. The USD price for file space would likely hover around the same price of something like S3. If the price paid by the user is too high, they are better of using S3 for storage. If the price paid to the user is too high, it creates an arbitrage situation in which someone can simply use S3 to host Filecoin data. This should theoretically provide more stability to the price of Filecoin in comparison to other virtual currencies which is simultaneously the biggest strength and weakness of most of these things.
- ErikHuisman 12y ago
  What if both the price paid by the user and thus paid to the miner is lower than S3, how would you arbitrage that? What would drive the price up to s3 prices? Scarcity?
  - slg 12y ago
    There wouldn't be a direct arbitrage opportunity if the price was lower, but we have to remember there is a reason why S3 has been so successful. HDD space and CPU cycles aren't free. Plus there is the redundancy issue that others have mentioned. Amazon can help address this because they control the servers and can nearly guarantee their uptime. Filecoin doesn't mention any requirement of uptime for its users, so it will likely need to combat the problem by just throwing numbers at it. We then have the user's costs (electricity, wear and tear on their machine, and opportunity costs of doing something else) multiplied by whatever number Filecoin feels is enough to guarantee redundancy. This is a long way of me saying I am not sure how much lower than S3 pricing they can realistically accomplish.
- anshukla 12y ago
  Filecoin providers may feasibly beat S3 on price in USD if (i) they're betting on the protocol because up-front losses may be more than compensated for in investment returns on Filecoin or (ii) the alternate for them is making no money on unused disk space.
  - slg 12y ago
jacob019 12y ago
Users pay to Put data into the network. Are there any ongoing payments for storage? I don't see any way to delete data either. If the network promises permanent free storage after a one time fee, it's sure to fail eventually.
- _prometheus 12y ago
  Yeah, the payment effectively buys storage for a certain amount of time. Read the discussion on Rewards here: http://filecoin.io/filecoin.pdf http://filecoin.io/filecoin.pdf
  - jacob019 12y ago
    "Transaction issuers must pay for all rewards up-front" Will there be a way to make ongoing payments for data retention? With a proper API this could be used for secure distributed backup, that would be a killer use and something I would gladly pay "miners" for.
    - rdrey 12y ago
      The way I understand it, Get'ing the data also publishes it to existing miners, so they can start storing it if it's a piece with a high reward value. The TTL is decreased by the Get operation, though. By understanding how valuable your data is you can choose a time to Put it again (while reward/TTL > market rate) with a new reward attached. Obviously that reward/TTL should be higher than the last one to buy you a bunch of safe Gets again.
orik 12y ago
Is this associated with http://filecoin.org/ http://filecoin.org/ (febuary 23th, 2014)?
jude- 12y ago
I'm surprised the whitepaper didn't cite PermaCoin [1], which addresses the same problem with proofs-of-retrievability and uses the BitCoin blockchain for storage. Would the authors like to comment on how their work differs? [1] http://cs.umd.edu/~amiller/permacoin.pdf http://cs.umd.edu/~amiller/permacoin.pdf
- _prometheus 12y ago
  Hey, yeah. I found out about permacoin (which is great!) after designing this construction. Turns out that permacoin and filecoin are really compatible, and complementary. amiller and I will be talking about it in the near future. (btw, permacoin's non-outsourceability is a great contribution! it should also be added to regular bitcoin) And, to answer your question, I didn't cite permacoin here because I would have to address a whole range of things like how they overlap/differ/which one is good for what parts. Perhaps I should've, but I think it's best left to a future paper (hinted at in the discussion). Stay tuned! :)
  - leni536 12y ago
    I hope you can take more hints from permacoin. There is an important economical point in permacoin's whitepaper: "The expected reward per unit of work is approximately constant, even for very small investments of computational effort. This prevents large participants from monopolizing the system and driving out ordinary participants." Maybe I don't have a deep understanding how filecoin works, but as far as I know if you don't have the storage capacity which can even reach the current difficulty then you have an expected reward of 0. Maybe it's a non-issue and the difficulty will never reach let's say 10MB, but if it reaches 1TB then it will be really hard to hop in the mining business.
- socrates1024 12y ago
  We designed Permacoin with the goal of being rigorously secure. Our design addressed a bunch of problems that aren't (yet) addressed by filecoin. - Erasure coding (over the whole dataset) guarantees that you can't just drop some of the filed and make them unrecoverable. - Preprocessing the entire dataset to be full entropy guarantees that it isn't cheaper to store one portion of the dataset than another (for example a segment with just all zeroes). - Outsource resistance: it would be a shame if someone just runs a cloud service to store the file, which would be more efficient than using local hard drives. Arguably this is already happening with Bitcoin cloud mining. Permacoin and the follow up work on nonoutsourceable puzzles address this. Without these features, the concern is that selfish miners might find ways to earn more filecoins without actually providing the useful service. Incentives are correctly aligned only when there's no way to game or cheat the system. Filecoin provides a richer interface (get and put for individual files) than Permacoin (which only guarantees recovery of the whole dataset at once). We would have supported this interface if we could figure out how to, while maintaining the same standard security properties! This is difficult (but hopefully possible) future work. On the other hand: - Maybe it will be a long time before practical ways to game filecoin catch on. - if filecoin builds a community that values altruistic participation like Tor (and even Bitcoin to some degree) it will be useful anyway - it could be possible to improve filecoin with features from Permacoin - there's a lot to learn from the actual launch of a system like this with real users
xiphias 12y ago
Is there a way to host a file for 1 month, 2 month, 1 year.... by paying more/less (and maybe increasing the storage time by additional payments). That would be the right model that could translate to S3 storage with the same price. Without a time component, storage prices have strange meanings.
- xiphias 12y ago
  Is the thinking behind infinite storage that the price of storage decreases exponentially so the cost of keeping a file forever is limited?
Taek 12y ago
Maybe I've missed something, but how does Filecoin handle redundancy, and how does Filecoin handle delegation? The devs seem to hope that by making the rare pieces valuable, they will somehow avoid having low redundancy. But what happens when everyone joins a large pool like Ghash.io, and every file has only a single point of failure? I would not trust my data to Filecoin, it seems too easy for files to get permanently lost. There are no guarantees that every file piece will be managed by at least N nodes.
- _prometheus 12y ago
  Great questions :) -- this borders into future announcements, so sadly I'll have to defer to the answer on this paper, which you summarize well here. (Not asking you to believe, only to wait a bit longer.) Also, this answer is actually not bad: information on the distribution of pieces will become widely available so users can tune rewards accordingly. Storage pools also have a ceiling: after backing up the entire piece set, it's better for them to compete than to keep growing.
  - Taek 12y ago
    After backing up the entire piece set, the requirement per-node for joining the pool declines - say 100GB each instead of 500GB each. Or perhaps you have a setup where the pool master takes care of storing all the data and the workers all do POW. Then it's up to the pool to tune prices so that workers can profit as much as possible. You might end up with multiple competing pools, but it puts the barrier-to-entry of being a new pool at a high price. Everyone will mine for the cheapest pool, and the 1 pool that can provide all the storage at the lowest cost is the 1 pool that will control all of the data on the network. It seems like a fragile solution.
    - _prometheus 12y ago
      In equilibrium, people will still compete to avoid sharing the reward with others if they can avoid it. You don't need a full pool to have high expected reward. Anyway, stay tuned ;) (+ for getting rid of PoW altogether).
jseip 12y ago
This is an amazing business model. If they can make it user-friendly (big if) they'll essentially be offering a more cost-effective, more robust alternative to Dropbox. This appears to be truly disintermediated (P2P) secure file sharing?
- fatrachet 12y ago
  Well Dropbox/Gdrive/Skydrive etc. all offer free storage in the range of 15-30gb, which is enough for most normal users. For professionell use there is S3,glacier, box etc. and they all offer known security and redundancy. Most businesses won't touch anything related to bitcoin and things lacking a reliable SLA. While a lot of people have unused storage space, many lack usable upload bandwidth and have low availability, so I would assume a lot of copies would be needed. It's gonna be very interesting to see if stuff like this or storj, which was posted earlier today, can compete with current services. But as someone with a lot of servers with unused space I sure hope so.
  - lukifer 12y ago
    One has to wonder how much unused disk space is sitting on the generic Dell boxes in various cubicle farms, libraries, etc. Could be decent trickle revenue, and it wouldn't impede usability/performance the way mining does.
    - _prometheus 12y ago
      Yes. There's so many computing resources going to waste. We're changing that.
  - _prometheus 12y ago
    Lots of businesses are already relying on Bitcoin for revenue. The reliability -- and other properties -- of a network like Filecoin can be measured easily. Businesses are all about numbers. :)
- justin_ 12y ago
  Unfortunately there's no guarantee that the files will stay around forever. While cheap and useful, I don't think it will be "robust". Freenet offers P2P file sharing as well, but much like this project there's no guarantee the data will survive for long.
bamdadd 12y ago
MaidSafe is doing similar kind of thing. I think they have much better plan and idea http://maidsafe.net/ http://maidsafe.net/ http://www.safecoin.io/ http://www.safecoin.io/
andrewchambers 12y ago
I always wanted proof of work to be computation power which is saved up over time, which can be spent as computation bursts on a network, or sold. Perhaps using pnacl binaries to to the computations. Rather than waste CPU power on mindless hashing, you could accomplish real computation tasks. This is pretty cool too though.
- _prometheus 12y ago
  are you interested in working on that? (useful computations) Or storage. Ping me at juan@benet.ai
- lukifer 12y ago
  "Proof of Non-Work" would be a very interesting concept, although I can't think of a way to do it. (In a sense, Proof of Work is an expensive method to accomplish Proof of Non-Work.)
bweitzman 12y ago
This is neat. I did some work on a data storage platform through a bittorrent tracker/community back in 2010. There was no concept of a currency instead relying on the tracker to prioritize downloads since nobody would have any incentive to download any particular file. The more you stored, the more you could store. It's neat to see this concept popping up now, http://storj.io/ http://storj.io/ is another platform that popped up recently.
JeremyBanks 12y ago
Very interesting. Are files expected to be stored forever, even if they're never requested?
- trevorcreech 12y ago
  Yes. Filecoin uses proof-of-retrievability i.e. you prove that you can access arbitrary parts of a file without needing to always provide the whole file.
  - JeremyBanks 12y ago
    Hmm. Does this mean that the cost of storing new data over time will grow, to support the larger and larger amount of existing data that needs to be stored?
chipsy 12y ago
I'm hugely optimistic about the market for "proof-of-application" cryptocurrency since it's also the next step in "race to the bottom" - make use of the spare resources of all our commodity electronics and eliminate a centralized authority simultaneously. We've had plenty of peer-to-peer systems, but the incentives are mostly wrong to make them scale up past the niches of the black market and political radicals.
em3rgent0rdr 12y ago
Unfortunately https not working, because they are using github pages. They should really put up a https site especially considering they are a crypto project and are soliciting email addresses. At least a TOR hidden service or a namecoined https site.
- _prometheus 12y ago
  Ack! Great point. We'll upgrade to https soon. Wish github supported https in pages :(
lappa 12y ago
Filecoin uses what they call "proof-of-retrievability". The thing about proof-of-work is that it is a proof. Assuming that sha256d is unpredictable, you will on average have use a certain amount of energy to find a low hash value. With proof-of-retrieval, the name itself is lying. There is no way to prove that a piece of data went from one person to another. It should be called invoking-trust-in-retrieval-and-proof-of-having-the-file. Along with that, because hashes have the property of being irreversible, blocks can be hashed and the winner cannot change the block without doing work again. Because of these properties, filecoin must use PoW (which their whitepaper states they do). According to their whitepaper, you must first construct the proof of ownership in the file, then you perform the PoW with that seed. This leads to huge economies of scale because the startup cost of mining at equilibrium will include storing an abundance of files (probably all of them).
- kaoD 12y ago
  > This leads to huge economies of scale because the startup cost of mining at equilibrium will include storing an abundance of files (probably all of them). I was following until that line. Why 'all of them'?
  - lappa 12y ago
    You make more money the more you host if I understand the whitepaper correctly.
- beejiu 12y ago
  > There is no way to prove that a piece of data went from one person to another. Actually, you can prove that a piece of data went from one person to another, but others in the network must also have a copy. I was thinking a few months back about applying a similar idea to Bit-Torrent / P2P sharing. It was possible to prove that the file was sent given that all nodes had a copy of the file; so the idea is more feasible for file sharing than it is personal file storage.
  - lappa 12y ago
    What you are referring to isn't a proof, it is invoking trust in a set of people.
pessimizer 12y ago
If the price lowers, I'm going to want to use it to store files instead of my hard drive at home, right? Can you share links with other people to fetch files? If the price gets higher, I'll want to buy more hard drives for more people to pay me to use my space. Is that how this works, or am I misreading? If it works like that, it's pure genius. You'd be able to get a lot of use out of it without ever converting it. This is a weird thought, but: could filecoin prop up the value of bitcoin, because the only way to convert filecoin to and from hard currency would be through bitcoin? edit: Could I boot from this? Could my phone boot from this? edit2: I'm assuming that the race to claim the proof reward will encourage low latency retrievals. Would I be able to tune my client in order to choose my own balance between latency and security, for example only rewarding quick responders but paying them highly to encourage a more low latency mix? Would that open this new cloud disk space market to quant trading strategies? edit3: If latency matters, then network topology matters. Different places on the network would be more lucrative than others. edit4: I'm seeing a workflow where I buy a hard drive, partition it into a local partition and a filecoin partition, then open it up to the network. After a few days, my filecoin partition is full, and I've generated a lot of filecoin. I use that filecoin to start copying files into a mounted virtual filecoin filesystem. I can rsync between local filesystems and filecoin filesystems. I could actually delegate all of that activity to a single fileserver at the edge of my home network, and have all of my home computers mount files from that (and phones and home automation could call home to it too.)
- pessimizer 12y ago
  By combining a filecoin client with a bitcoin client, you could also run a decentralized filecoin-bitcoin exchange market.
- rakoo 12y ago
  I guess the important aspects are: - Filecoin is not a filesystem. The whitepaper doesn't even talk about the technical implementations of storage. Filecoin focusses on "how to retribute peers" more than "how to store data". - Filecoin is supposed to be used on the network - Filecoin is supposed to be used in a cluster of mutually distrusting peers, who by the power of the oldest and most effficient incentive ever (aka money) can come to an agreement that is beneficial to everyone. Given these, I wouldn't advise you replace your hard drives with it, not even your home computers; high latency should be expected. OTOH, Filecoin is the way to go if you want to backup your whole TBs of data without paying thousands of dollars each month. Note: I only speak from a very quick glance at the whitepaper. When I see something like that I always rush to the technical explanations, and I was a little bit sad that it only talked about the coin aspect...
xur17 12y ago
How is bandwidth handled / incentivized? Is this intended for long term storage (few retrievals), or for files that are accessed frequently?
- _prometheus 12y ago
  more for long term storage. see ipfs for short term.
nullc 12y ago
This was discussed in #bitcoin-wizards recently: http://download.wpsoftware.net/bitcoin/wizards/2014-07-17.html http://download.wpsoftware.net/bitcoin/wizards/2014-07-17.ht...
- wyager 12y ago
  So it sounds like the incentive model is kind of fundamentally broken. Is it correct to say that miners are incentivized to A) store trash data in a way that lets them mine with zero actual storage backing it up and B) consolidate their storage into the same hard drive, until there exists only a single copy of the data on the entire network?
  - nullc 12y ago
    Yes. Their counter argument to (A) that the trash data will cost more to store than the user gets for retrieving it (though this isn't true if the attack is successful and results in them being the only miner/retrieval server), and (B) they don't seem to consider a problem. There are other issues— e.g. arbitrary data can be censored and the only harm is lost mining revenue proportional to the censored data. (e.g. you could censor 0.1% of all data, and probably make every file unreadable (just by censoring the first word of it) but only be unable to answer 0.1% of the file queries). I don't really see the point— having a file storage service that you can pay for with cryptocurrency make sense, but integrating it in this manner seems to serve no technical purpose other than to fuel investment speculation and otherwise just makes the system worse.
sp4ke 12y ago
There's also http://storj.io/ http://storj.io/ and they are crowdfunding the project over the decentralized financial exchange counterparty like swarm :-)
- nbody 12y ago
  They raised 450 BTC within a day, seems they have a following. Hopefully one of them will get this up and running soon.
baumbart 12y ago
Ok, I see some similarities to other projects here. Let's introduce one additional layer of abstraction that we call "resources-coin". Disk space is not the only limited resource that can be shared using an economic system; CPU cycles (not for hashing, but doing actual work), network bandwidth, storage space (like this project) are all things that have some value that one can be rewarded for. I bet there are many more ways. The difference with traditional bitcoin is that, we're not talking about an absolutely limited resource here. While bitcoin will stop growing at a certain point, no matter what happens, storage space on earth will always grow. So will the overall bandwidth, overall computational power etc. And that is why these things are not going to work on the long term. You cannot build a currency based on reward.
- andruby 12y ago
  Can you elaborate on why you cannot build a currency based on reward?
- kylebrown 12y ago
  These things will work, they just won't be good for "store of value" like bitcoin. Supply of disk space, bandwidth etc keeps growing, but the demand for those resources grows as well. Hard to predict the ratio of supply to demand, and therefore the price, but I think they could work very well in the long term. Not as a universal payment currency, but rather as a token (with fluctuating price) for some particular type of service. I see the future of "resource coins" as being efficient mechanisms for p2p resource distribution. Conventionally, a customer would use their credit card to pay some cloud hosting company. The conventional hosting company is a business which operates as a privately managed bureaucracy. A P2P coin operates as standard protocol acting as a meta-company, aka a DAC/DAO - Digital Autonomous Company/Organization. In a conventional business, a private bureaucracy manages a resource uses it to provide a service (bandwidth, disk space, and CPU as managed by a hosting company). P2P coins replace that centrally-operated middle man with a decentralized algorithm and account database. Bitcoin is like a meta-company which operates a distributed ledger internet payments. Why stop at payments? If its efficient, we should *coin all the things.
  - 12y ago
doctorKrieger 12y ago
datacoin anyone?
sly_g 12y ago
Will your reward depend on the volume you dedicated to store other people's files?
z3t4 12y ago
What I think would be a limited factor is the bandwidth: You need at least 35 megabit per second for streaming full HD video+sound ... While most people are stuck on 5 Mbit Internet or 50 GB/month quota.
ff_ 12y ago
That's a beautiful thing, I was thinking about building a system like that, but without the cryptocurrency thing, it's genial. Reading the whitepaper one thing isn't clear: will the pieces of data be encrypted before sending them to the network?
bascule 12y ago
All that is old is new again: https://en.wikipedia.org/wiki/Mnet_(peer-to-peer_network) https://en.wikipedia.org/wiki/Mnet_(peer-to-peer_network)

281 points by _prometheus 12y ago

Is this one also vulnerable to people creating fake nodes that "store files" for each other to earn coins for doing nothing?

dutchbrit 12y ago
Exactly my thought! Seems like a big flaw.
bdcs 12y ago
The authors address Sybil attacks in the paper. Essentially, there is a PoW style consensus using hashes of files+metadata (as opposed to bitcoin which is hashes of nothing + metadata)
- abc123xyz 12y ago
  whats to stop the node replying with the hash of the file saying "yes the file is here" when in reality no file is there
  - ZitchDog 12y ago
    Haven't read the paper, but could you ask for a hash of a random offset+byte count of the file?
    - slipstream- 12y ago
      Exactly what I was thinking.
    - wyager 12y ago
      That only works on direct person-to-person systems, not decentralized systems where the entire network has to reach consensus.
  - _prometheus 12y ago
    the proof-of-retrievability. see http://cseweb.ucsd.edu/~hovav/dist/verstore.pdf http://cseweb.ucsd.edu/~hovav/dist/verstore.pdf
    - philsnow 12y ago
      I'm at work and don't want to read either of the papers at the moment, but if A asks B to provide proof of having stored some bytes, could B proxy that request to C and replay the reply to A ?
      - justin_ 12y ago
        As long as a node you control has the data and can retrieve the file, I don't think there's a problem. If you could proxy the "proof of retrieval" than you could surely proxy the retrieval itself. Pooling resources shouldn't break anything.
        sp332 12y ago
        But you could make it look like several computers have copies of the data when there's really only one copy of the data. You would get credit/coins for multiple copies and reduce the redundancy of the files.
        anonymousDan 12y ago
        I haven't RTFA or any of the papers but couldn't you perhaps add a different nonce to each copy of the file you want stored before encrypting?
        justin_ 12y ago
        From what I read in the paper, it sounds like you _don't_ get paid for providing the data. You get paid for proving you have it when mining a block. Distributing your storage across multiples nodes would simply make it easier to "mine", but wouldn't get coins passively. That is, if I'm understanding the protocol correctly.
_prometheus 12y ago
No, files are contributed to the network as a whole, must be distributed, and require payment. See the discussions on mining and on rewards in the paper: http://filecoin.io/filecoin.pdf http://filecoin.io/filecoin.pdf
- JVIDEL 12y ago
  Hey sorry for the OT but shouldn't this be on the new Show section? just saying...
  - _prometheus 12y ago
    Well, it's not something you can TRY yet -- i think those are the guidelines? :)

flixic 12y ago

How to create new crypto-currency: - Pick something to base you currency on. Bit-, Name-, and File- are taken, but some might be left. - First letter, two stripes, here's a logo! - Publish a whitepaper. You must use LaTeX to typeset it. - No names or credit anywhere. Stay anonymous. Joking aside, this looks like something I will try. Have a spare 1TB connected to fast internet, and assume many others have as well.

samirmenon 12y ago
The Whois lookup of filecoin.io gives the name "Juan Batiz-Benet", so this isn't very anonymous.
- _prometheus 12y ago
  nope, it isn't ;)
  - tinco 12y ago
    So why isn't your name on the paper? (not judging, just curious)
_prometheus 12y ago
Haha! Well, the crypto and game theory took a very long time to get right. (and more of it to come!) Send feedback!

jothi 12y ago

How different is it from Storjcoin ?

mappum 12y ago
Storj doesn't make any technical innovations, their plan is essentially "make some nice looking web apps and let other people worry about inventing the things we need". Storjcoin is a standard blockchain, but actually isn't even neccessary at all. Filecoin actually addresses the technical challenges in building something like this, and maybe Storj will end up using it in their stack.
- ard10e 12y ago
  I spoke with super3, the Storj founder earlier today, and he was going over how they'll go about providing Proof of Storage. I'm not qualified to explain it myself, but expect a revised whitepaper in the next couple weeks that goes over it in detail.
_prometheus 12y ago
At first glance the concepts may seem similar, but this is a different thing altogether. Storj is a service layered on top of other blockchains. Filecoin is a new blockchain protocol itself, with its own contributions: mining construction, use of proof-of-retrievability, put/get as transactions, and more. Read: http://filecoin.io/filecoin.pdf http://filecoin.io/filecoin.pdf Soon will have simpler explanation. :)
abc123xyz 12y ago
Both smell like a scam, "coincidentally" today storj started fundraising by getting people to send them bitcoins for useless storjcoins they created, the worst part is people falling for such a scam, then again there's a fool born every minute which is a business model a lot of companies rely on. In this case and in storjcoin case the actors behind it all remain anonymous
- _prometheus 12y ago
  I can understand why -- with all the scams out there -- that would be one reaction. But note that Filecoin makes important technical contributions on top of the Bitcoin protocol. You can read the paper here: http://filecoin.io/filecoin.pdf http://filecoin.io/filecoin.pdf FWIW, we aren't really anonymous. Neither in the storj case (you can find who they are in their website), nor in the Filecoin case. (hi!)

possibilistic 12y ago

Unrelated to Filecoin, this brought to mind a question about the Bitcoin blockchain I've been pondering recently. Since users can put small arbitrary byte data in a transaction that gets included in the block chain (AFAIK), isn't it possible that an attacker could insert illicit / illegal data that everyone using Bitcoin replicates? Copyrighted materials, sensitive personal information such as SSNs, and moreover, child pornograpy immediately come to mind as things that could forever be persisted. If this is possible, wouldn't that make everyone using the bitcoin blockchain criminally liable for storing and/or transmitting that information? (In the US and similar jurisdictions.) I assume lightweight clients that only store a hash would be unaffected. But still, this seems like a serious issue.

wyager 12y ago
This has already been done many times.
_prometheus 12y ago
Yes, it's been known for a while that people can upload illegal bits into blockchains. Don't think transaction steganography would hold up in court though. You could do similar things with DNS records, etc. File storage services though have the added concern that they're specifically designed to distribute files. Filecoin addresses this by (a) encouraging everyone to pre-encrypt before adding, and (b) creating a fluid market based on incentives. Nodes aren't required to store every file, only strongly incentivized to do so. Nodes are free to follow strong counter incentives (say, a well-known list of "illegal hashes" published by authorities). In reality, the way many courts handle things like this is to target the entry points: the websites advertising which are the illegal hashes and how to get them. (e.g. the piratebay instead of bittorrent inc.)
- kang 12y ago
  Make encryption at source compulsory (since the data is uploaded through your software) and no need to tell the user about it. rather advertise it as a feature.

dubcanada 12y ago

Does "Renting disk space" not sound extremely sketchy to anyone else?

_prometheus 12y ago
Does "Renting your spare room" (airbnb) or "Renting your car" (uber) sound sketchy? It's all about how the protocol works and the security guarantees on top. Users should store pre-encrypted files. (except for public goods, like wikipedia)

Hilyin 12y ago

What about being liable for storing illegal files someone uploads to your filebox?

elux 12y ago
49310 83597 02850 19002 75777 67239 07649 57284 90777 21502 08632 08075 01840 97926 27885 09765 88645 57802 01366 00732 86795 44734 11283 17353 67831 20155 75359 81978 54505 48115 71939 34587 73300 38009 93261 95058 76452 50238 20408 11018 98850 42615 17657 99417 04250 88903 70291 19015 87003 04794 32826 07382 14695 41570 33022 79875 57681 89560 16240 30064 11151 69008 72879 83819 42582 71674 56477 48166 84347 92846 45809 29131 53186 00700 10043 35318 93631 93439 12948 60445 03709 91980 04770 94629 21558 18071 11691 53031 87628 84778 78354 15759 32891 09329 54473 50881 88246 54950 60005 01900 62747 05305 38116 42782 94267 47485 34965 25745 36815 11706 55028 19055 52656 22135 31463 10421 00866 28679 71144 46706 36692 19825 86158 11125 15556 50481 34207 68673 23407 65505 48591 08269 56266 69306 62367 99702 10481 23965 62518 00681 83236 53959 34839 56753 57557 53246 19023 48106 47009 87753 02795 61868 92925 38069 33052 04238 14996 99454 56945 77413 83356 89906 00587 08321 81270 48611 33682 02651 59051 66351 87402 90181 97693 93767 78529 28722 10955 04129 25792 57381 86605 84501 50552 50274 99477 18831 29310 45769 80909 15304 61335 94190 30258 81320 59322 77444 38525 50466 77902 45186 97062 62778 88919 79580 42306 57506 15669 83469 56177 97879 65920 16440 51939 96071 69811 12615 19561 02762 83233 98257 91423 32172 69614 43744 38105 64855 29348 87634 92103 09887 02878 74532 33132 53212 26786 33283 70279 25099 74996 94887 75936 91591 76445 88032 71838 47402 35933 02037 48885 06755 70658 79194 61134 19323 07814 85443 64543 75113 20709 86063 90746 41756 41216 35042 38800 29678 08558 67037 03875 09410 76982 11837 65499 20520 43682 55854 64228 85024 29963 32268 53691 24648 55000 75591 66402 47292 40716 45072 53196 74499 95294 48434 74190 21077 29606 82055 81309 23626 83798 79519 66199 79828 55258 87161 09613 65617 80745 66159 24886 60889 81645 68541 72136 29208 46656 27913 14784 66791 55096 51543 10113 53858 62081 96875 83688 35955 77893 91454 53935 68199 60988 08540 47659 07358 97289 89834 25047 12891 84162 65878 96821 85380 87956 27903 99786 29449 39760 54675 34821 25675 01215 17082 73710 76462 70712 46753 21024 83678 15940 00875 05452 54353 7. ^ https://en.wikipedia.org/wiki/Illegal_prime https://en.wikipedia.org/wiki/Illegal_prime Now what?
- TomGullen 12y ago

slg 12y ago

This is pretty interesting from a economic standpoint. The USD price for file space would likely hover around the same price of something like S3. If the price paid by the user is too high, they are better of using S3 for storage. If the price paid to the user is too high, it creates an arbitrage situation in which someone can simply use S3 to host Filecoin data. This should theoretically provide more stability to the price of Filecoin in comparison to other virtual currencies which is simultaneously the biggest strength and weakness of most of these things.

ErikHuisman 12y ago
What if both the price paid by the user and thus paid to the miner is lower than S3, how would you arbitrage that? What would drive the price up to s3 prices? Scarcity?
- slg 12y ago
  There wouldn't be a direct arbitrage opportunity if the price was lower, but we have to remember there is a reason why S3 has been so successful. HDD space and CPU cycles aren't free. Plus there is the redundancy issue that others have mentioned. Amazon can help address this because they control the servers and can nearly guarantee their uptime. Filecoin doesn't mention any requirement of uptime for its users, so it will likely need to combat the problem by just throwing numbers at it. We then have the user's costs (electricity, wear and tear on their machine, and opportunity costs of doing something else) multiplied by whatever number Filecoin feels is enough to guarantee redundancy. This is a long way of me saying I am not sure how much lower than S3 pricing they can realistically accomplish.
anshukla 12y ago
Filecoin providers may feasibly beat S3 on price in USD if (i) they're betting on the protocol because up-front losses may be more than compensated for in investment returns on Filecoin or (ii) the alternate for them is making no money on unused disk space.
- slg 12y ago

jacob019 12y ago

Users pay to Put data into the network. Are there any ongoing payments for storage? I don't see any way to delete data either. If the network promises permanent free storage after a one time fee, it's sure to fail eventually.

_prometheus 12y ago
Yeah, the payment effectively buys storage for a certain amount of time. Read the discussion on Rewards here: http://filecoin.io/filecoin.pdf http://filecoin.io/filecoin.pdf
- jacob019 12y ago
  "Transaction issuers must pay for all rewards up-front" Will there be a way to make ongoing payments for data retention? With a proper API this could be used for secure distributed backup, that would be a killer use and something I would gladly pay "miners" for.
  - rdrey 12y ago
    The way I understand it, Get'ing the data also publishes it to existing miners, so they can start storing it if it's a piece with a high reward value. The TTL is decreased by the Get operation, though. By understanding how valuable your data is you can choose a time to Put it again (while reward/TTL > market rate) with a new reward attached. Obviously that reward/TTL should be higher than the last one to buy you a bunch of safe Gets again.

orik 12y ago

Is this associated with http://filecoin.org/ http://filecoin.org/ (febuary 23th, 2014)?

jude- 12y ago

I'm surprised the whitepaper didn't cite PermaCoin [1], which addresses the same problem with proofs-of-retrievability and uses the BitCoin blockchain for storage. Would the authors like to comment on how their work differs? [1] http://cs.umd.edu/~amiller/permacoin.pdf http://cs.umd.edu/~amiller/permacoin.pdf

_prometheus 12y ago
Hey, yeah. I found out about permacoin (which is great!) after designing this construction. Turns out that permacoin and filecoin are really compatible, and complementary. amiller and I will be talking about it in the near future. (btw, permacoin's non-outsourceability is a great contribution! it should also be added to regular bitcoin) And, to answer your question, I didn't cite permacoin here because I would have to address a whole range of things like how they overlap/differ/which one is good for what parts. Perhaps I should've, but I think it's best left to a future paper (hinted at in the discussion). Stay tuned! :)
- leni536 12y ago
  I hope you can take more hints from permacoin. There is an important economical point in permacoin's whitepaper: "The expected reward per unit of work is approximately constant, even for very small investments of computational effort. This prevents large participants from monopolizing the system and driving out ordinary participants." Maybe I don't have a deep understanding how filecoin works, but as far as I know if you don't have the storage capacity which can even reach the current difficulty then you have an expected reward of 0. Maybe it's a non-issue and the difficulty will never reach let's say 10MB, but if it reaches 1TB then it will be really hard to hop in the mining business.
socrates1024 12y ago
We designed Permacoin with the goal of being rigorously secure. Our design addressed a bunch of problems that aren't (yet) addressed by filecoin. - Erasure coding (over the whole dataset) guarantees that you can't just drop some of the filed and make them unrecoverable. - Preprocessing the entire dataset to be full entropy guarantees that it isn't cheaper to store one portion of the dataset than another (for example a segment with just all zeroes). - Outsource resistance: it would be a shame if someone just runs a cloud service to store the file, which would be more efficient than using local hard drives. Arguably this is already happening with Bitcoin cloud mining. Permacoin and the follow up work on nonoutsourceable puzzles address this. Without these features, the concern is that selfish miners might find ways to earn more filecoins without actually providing the useful service. Incentives are correctly aligned only when there's no way to game or cheat the system. Filecoin provides a richer interface (get and put for individual files) than Permacoin (which only guarantees recovery of the whole dataset at once). We would have supported this interface if we could figure out how to, while maintaining the same standard security properties! This is difficult (but hopefully possible) future work. On the other hand: - Maybe it will be a long time before practical ways to game filecoin catch on. - if filecoin builds a community that values altruistic participation like Tor (and even Bitcoin to some degree) it will be useful anyway - it could be possible to improve filecoin with features from Permacoin - there's a lot to learn from the actual launch of a system like this with real users

xiphias 12y ago

Is there a way to host a file for 1 month, 2 month, 1 year.... by paying more/less (and maybe increasing the storage time by additional payments). That would be the right model that could translate to S3 storage with the same price. Without a time component, storage prices have strange meanings.

xiphias 12y ago
Is the thinking behind infinite storage that the price of storage decreases exponentially so the cost of keeping a file forever is limited?

Taek 12y ago

Maybe I've missed something, but how does Filecoin handle redundancy, and how does Filecoin handle delegation? The devs seem to hope that by making the rare pieces valuable, they will somehow avoid having low redundancy. But what happens when everyone joins a large pool like Ghash.io, and every file has only a single point of failure? I would not trust my data to Filecoin, it seems too easy for files to get permanently lost. There are no guarantees that every file piece will be managed by at least N nodes.

_prometheus 12y ago
Great questions :) -- this borders into future announcements, so sadly I'll have to defer to the answer on this paper, which you summarize well here. (Not asking you to believe, only to wait a bit longer.) Also, this answer is actually not bad: information on the distribution of pieces will become widely available so users can tune rewards accordingly. Storage pools also have a ceiling: after backing up the entire piece set, it's better for them to compete than to keep growing.
- Taek 12y ago
  After backing up the entire piece set, the requirement per-node for joining the pool declines - say 100GB each instead of 500GB each. Or perhaps you have a setup where the pool master takes care of storing all the data and the workers all do POW. Then it's up to the pool to tune prices so that workers can profit as much as possible. You might end up with multiple competing pools, but it puts the barrier-to-entry of being a new pool at a high price. Everyone will mine for the cheapest pool, and the 1 pool that can provide all the storage at the lowest cost is the 1 pool that will control all of the data on the network. It seems like a fragile solution.
  - _prometheus 12y ago
    In equilibrium, people will still compete to avoid sharing the reward with others if they can avoid it. You don't need a full pool to have high expected reward. Anyway, stay tuned ;) (+ for getting rid of PoW altogether).

jseip 12y ago

This is an amazing business model. If they can make it user-friendly (big if) they'll essentially be offering a more cost-effective, more robust alternative to Dropbox. This appears to be truly disintermediated (P2P) secure file sharing?

fatrachet 12y ago
Well Dropbox/Gdrive/Skydrive etc. all offer free storage in the range of 15-30gb, which is enough for most normal users. For professionell use there is S3,glacier, box etc. and they all offer known security and redundancy. Most businesses won't touch anything related to bitcoin and things lacking a reliable SLA. While a lot of people have unused storage space, many lack usable upload bandwidth and have low availability, so I would assume a lot of copies would be needed. It's gonna be very interesting to see if stuff like this or storj, which was posted earlier today, can compete with current services. But as someone with a lot of servers with unused space I sure hope so.
- lukifer 12y ago
  One has to wonder how much unused disk space is sitting on the generic Dell boxes in various cubicle farms, libraries, etc. Could be decent trickle revenue, and it wouldn't impede usability/performance the way mining does.
  - _prometheus 12y ago
    Yes. There's so many computing resources going to waste. We're changing that.
- _prometheus 12y ago
  Lots of businesses are already relying on Bitcoin for revenue. The reliability -- and other properties -- of a network like Filecoin can be measured easily. Businesses are all about numbers. :)
justin_ 12y ago
Unfortunately there's no guarantee that the files will stay around forever. While cheap and useful, I don't think it will be "robust". Freenet offers P2P file sharing as well, but much like this project there's no guarantee the data will survive for long.

bamdadd 12y ago

MaidSafe is doing similar kind of thing. I think they have much better plan and idea http://maidsafe.net/ http://maidsafe.net/ http://www.safecoin.io/ http://www.safecoin.io/

andrewchambers 12y ago

I always wanted proof of work to be computation power which is saved up over time, which can be spent as computation bursts on a network, or sold. Perhaps using pnacl binaries to to the computations. Rather than waste CPU power on mindless hashing, you could accomplish real computation tasks. This is pretty cool too though.

_prometheus 12y ago
are you interested in working on that? (useful computations) Or storage. Ping me at juan@benet.ai
lukifer 12y ago
"Proof of Non-Work" would be a very interesting concept, although I can't think of a way to do it. (In a sense, Proof of Work is an expensive method to accomplish Proof of Non-Work.)

bweitzman 12y ago

This is neat. I did some work on a data storage platform through a bittorrent tracker/community back in 2010. There was no concept of a currency instead relying on the tracker to prioritize downloads since nobody would have any incentive to download any particular file. The more you stored, the more you could store. It's neat to see this concept popping up now, http://storj.io/ http://storj.io/ is another platform that popped up recently.

JeremyBanks 12y ago

Very interesting. Are files expected to be stored forever, even if they're never requested?

trevorcreech 12y ago
Yes. Filecoin uses proof-of-retrievability i.e. you prove that you can access arbitrary parts of a file without needing to always provide the whole file.
- JeremyBanks 12y ago
  Hmm. Does this mean that the cost of storing new data over time will grow, to support the larger and larger amount of existing data that needs to be stored?

chipsy 12y ago

I'm hugely optimistic about the market for "proof-of-application" cryptocurrency since it's also the next step in "race to the bottom" - make use of the spare resources of all our commodity electronics and eliminate a centralized authority simultaneously. We've had plenty of peer-to-peer systems, but the incentives are mostly wrong to make them scale up past the niches of the black market and political radicals.

em3rgent0rdr 12y ago

Unfortunately https not working, because they are using github pages. They should really put up a https site especially considering they are a crypto project and are soliciting email addresses. At least a TOR hidden service or a namecoined https site.

_prometheus 12y ago
Ack! Great point. We'll upgrade to https soon. Wish github supported https in pages :(

lappa 12y ago

Filecoin uses what they call "proof-of-retrievability". The thing about proof-of-work is that it is a proof. Assuming that sha256d is unpredictable, you will on average have use a certain amount of energy to find a low hash value. With proof-of-retrieval, the name itself is lying. There is no way to prove that a piece of data went from one person to another. It should be called invoking-trust-in-retrieval-and-proof-of-having-the-file. Along with that, because hashes have the property of being irreversible, blocks can be hashed and the winner cannot change the block without doing work again. Because of these properties, filecoin must use PoW (which their whitepaper states they do). According to their whitepaper, you must first construct the proof of ownership in the file, then you perform the PoW with that seed. This leads to huge economies of scale because the startup cost of mining at equilibrium will include storing an abundance of files (probably all of them).

kaoD 12y ago
> This leads to huge economies of scale because the startup cost of mining at equilibrium will include storing an abundance of files (probably all of them). I was following until that line. Why 'all of them'?
- lappa 12y ago
  You make more money the more you host if I understand the whitepaper correctly.
beejiu 12y ago
> There is no way to prove that a piece of data went from one person to another. Actually, you can prove that a piece of data went from one person to another, but others in the network must also have a copy. I was thinking a few months back about applying a similar idea to Bit-Torrent / P2P sharing. It was possible to prove that the file was sent given that all nodes had a copy of the file; so the idea is more feasible for file sharing than it is personal file storage.
- lappa 12y ago
  What you are referring to isn't a proof, it is invoking trust in a set of people.

pessimizer 12y ago

If the price lowers, I'm going to want to use it to store files instead of my hard drive at home, right? Can you share links with other people to fetch files? If the price gets higher, I'll want to buy more hard drives for more people to pay me to use my space. Is that how this works, or am I misreading? If it works like that, it's pure genius. You'd be able to get a lot of use out of it without ever converting it. This is a weird thought, but: could filecoin prop up the value of bitcoin, because the only way to convert filecoin to and from hard currency would be through bitcoin? edit: Could I boot from this? Could my phone boot from this? edit2: I'm assuming that the race to claim the proof reward will encourage low latency retrievals. Would I be able to tune my client in order to choose my own balance between latency and security, for example only rewarding quick responders but paying them highly to encourage a more low latency mix? Would that open this new cloud disk space market to quant trading strategies? edit3: If latency matters, then network topology matters. Different places on the network would be more lucrative than others. edit4: I'm seeing a workflow where I buy a hard drive, partition it into a local partition and a filecoin partition, then open it up to the network. After a few days, my filecoin partition is full, and I've generated a lot of filecoin. I use that filecoin to start copying files into a mounted virtual filecoin filesystem. I can rsync between local filesystems and filecoin filesystems. I could actually delegate all of that activity to a single fileserver at the edge of my home network, and have all of my home computers mount files from that (and phones and home automation could call home to it too.)

pessimizer 12y ago
By combining a filecoin client with a bitcoin client, you could also run a decentralized filecoin-bitcoin exchange market.
rakoo 12y ago
I guess the important aspects are: - Filecoin is not a filesystem. The whitepaper doesn't even talk about the technical implementations of storage. Filecoin focusses on "how to retribute peers" more than "how to store data". - Filecoin is supposed to be used on the network - Filecoin is supposed to be used in a cluster of mutually distrusting peers, who by the power of the oldest and most effficient incentive ever (aka money) can come to an agreement that is beneficial to everyone. Given these, I wouldn't advise you replace your hard drives with it, not even your home computers; high latency should be expected. OTOH, Filecoin is the way to go if you want to backup your whole TBs of data without paying thousands of dollars each month. Note: I only speak from a very quick glance at the whitepaper. When I see something like that I always rush to the technical explanations, and I was a little bit sad that it only talked about the coin aspect...

xur17 12y ago

How is bandwidth handled / incentivized? Is this intended for long term storage (few retrievals), or for files that are accessed frequently?

_prometheus 12y ago
more for long term storage. see ipfs for short term.

nullc 12y ago

This was discussed in #bitcoin-wizards recently: http://download.wpsoftware.net/bitcoin/wizards/2014-07-17.html http://download.wpsoftware.net/bitcoin/wizards/2014-07-17.ht...

wyager 12y ago
So it sounds like the incentive model is kind of fundamentally broken. Is it correct to say that miners are incentivized to A) store trash data in a way that lets them mine with zero actual storage backing it up and B) consolidate their storage into the same hard drive, until there exists only a single copy of the data on the entire network?
- nullc 12y ago
  Yes. Their counter argument to (A) that the trash data will cost more to store than the user gets for retrieving it (though this isn't true if the attack is successful and results in them being the only miner/retrieval server), and (B) they don't seem to consider a problem. There are other issues— e.g. arbitrary data can be censored and the only harm is lost mining revenue proportional to the censored data. (e.g. you could censor 0.1% of all data, and probably make every file unreadable (just by censoring the first word of it) but only be unable to answer 0.1% of the file queries). I don't really see the point— having a file storage service that you can pay for with cryptocurrency make sense, but integrating it in this manner seems to serve no technical purpose other than to fuel investment speculation and otherwise just makes the system worse.

sp4ke 12y ago

There's also http://storj.io/ http://storj.io/ and they are crowdfunding the project over the decentralized financial exchange counterparty like swarm :-)

nbody 12y ago
They raised 450 BTC within a day, seems they have a following. Hopefully one of them will get this up and running soon.

baumbart 12y ago

Ok, I see some similarities to other projects here. Let's introduce one additional layer of abstraction that we call "resources-coin". Disk space is not the only limited resource that can be shared using an economic system; CPU cycles (not for hashing, but doing actual work), network bandwidth, storage space (like this project) are all things that have some value that one can be rewarded for. I bet there are many more ways. The difference with traditional bitcoin is that, we're not talking about an absolutely limited resource here. While bitcoin will stop growing at a certain point, no matter what happens, storage space on earth will always grow. So will the overall bandwidth, overall computational power etc. And that is why these things are not going to work on the long term. You cannot build a currency based on reward.

andruby 12y ago
Can you elaborate on why you cannot build a currency based on reward?
kylebrown 12y ago
These things will work, they just won't be good for "store of value" like bitcoin. Supply of disk space, bandwidth etc keeps growing, but the demand for those resources grows as well. Hard to predict the ratio of supply to demand, and therefore the price, but I think they could work very well in the long term. Not as a universal payment currency, but rather as a token (with fluctuating price) for some particular type of service. I see the future of "resource coins" as being efficient mechanisms for p2p resource distribution. Conventionally, a customer would use their credit card to pay some cloud hosting company. The conventional hosting company is a business which operates as a privately managed bureaucracy. A P2P coin operates as standard protocol acting as a meta-company, aka a DAC/DAO - Digital Autonomous Company/Organization. In a conventional business, a private bureaucracy manages a resource uses it to provide a service (bandwidth, disk space, and CPU as managed by a hosting company). P2P coins replace that centrally-operated middle man with a decentralized algorithm and account database. Bitcoin is like a meta-company which operates a distributed ledger internet payments. Why stop at payments? If its efficient, we should *coin all the things.
- 12y ago

doctorKrieger 12y ago

datacoin anyone?

sly_g 12y ago

Will your reward depend on the volume you dedicated to store other people's files?

z3t4 12y ago

What I think would be a limited factor is the bandwidth: You need at least 35 megabit per second for streaming full HD video+sound ... While most people are stuck on 5 Mbit Internet or 50 GB/month quota.

ff_ 12y ago

That's a beautiful thing, I was thinking about building a system like that, but without the cryptocurrency thing, it's genial. Reading the whitepaper one thing isn't clear: will the pieces of data be encrypted before sending them to the network?

bascule 12y ago

All that is old is new again: https://en.wikipedia.org/wiki/Mnet_(peer-to-peer_network) https://en.wikipedia.org/wiki/Mnet_(peer-to-peer_network)