Building Ripple on Ethereum?

How feasible would it be to construct a Ripple clone on Ethereum?

Comments

  • levilevi Member Posts: 11
    If I am not mistaken, there is an Ethereum clone on Ripple.

    https://ripple.com/tag/smart-contracts/
  • StephanTualStephanTual London, EnglandMember, Moderator Posts: 1,282 mod
    @levi, you're mistaken, Codius is not an 'Ethereum clone' but an oracle platform, and it doesn't 'run on Ripple' either, it's just funded by them.
  • levilevi Member Posts: 11
    Thank you for the clarification.
  • eric22eric22 Member Posts: 12
    edited May 2015
    @celticwarrior72 totally feasible. ethereum is a decentralized application platform, and ripple is one application that I'd choose to run on it.

    Ryan Fugger's invention belongs on the Ethereum web 3.0 paradigm. Ripplelabs did a great job and used the technology of their time. Ethereum often claims to be pioneering the 'gas' idea, but Ripplelabs pioneered it two years ago with XRP as a gas. Ripplelabs gave us alot of new concepts. I'm sure Ryan, and Jed McKaleb, and the other pioneers of the ripple-framework knew that disruptive innovations like the EVM were coming, and they just went along with what they had access to at the time. Ethereums blockchain 2.0 is a platform technology, and Ripple is content, and whoever wants to continue Ryan's legacy will probably choose to use the best platform available, which some months from now might be Vitaliks Ethereum.
  • celticwarrior72celticwarrior72 Member Posts: 83
    @eric22 I think the potential applications of Ethereum within the financial are many and quite interesting
  • ThePiachuThePiachu Member Posts: 16
    It is possible, but the cost might be high and the network might be slow. Ripple is a compiled, optimized code. Replacing that with a script won't be nearly as efficient.
  • eric22eric22 Member Posts: 12
    oh, Stellar seems to be on Horizon already, https://github.com/stellar/horizon
  • markjenkinsmarkjenkins Member Posts: 1
    You can build something like Counterparty or maybe even something like BitShares on top of Ethereum (harder problem), but Ripple is a different beast.

    Don't think of Ripple as simply a place to record counter party relationships (debts). Their ledger allows you to make payments that depend on a chain reaction of balance trading through many, many counterparties. That's the rippling.

    Such transactions do not work as well on a blockchain. You need a network that can come to a truly finalized concensus very quickly to make rippling work. The more rippling your transaction requires the greater the chance that someone else's transaction is going to be in conflict in terms of which chains of counter party liquidity is available to make make the transaction possible.

    User's expect that the process of making a payment is a one time action that requires no follow up or waiting to see if the transaction is going through. It's perfectly fine if the receiver is stuck waiting for confirmation, but for sender it has to be pay and walk away.

    A blockchain is able to deliver this sending user experience if the user is directly moving a token from themselves to another party. All the user has to done is include a high enough transaction fee and not double spend.

    But if Alice is paying Bob through several intermediate chains of Carols, there's no way for Alice to just send payment and walk away. There's always the risk that one of the intermediate Carols will no longer have the liquidity Alice thought they had.

    Short block times combined with Alice waiting for the first block doesn't solve this problem. if Alice is directly sending a token to Bob then only Bob has to worry about Alice having enough hashing power to double spend through an alternative block chain so Bob just waits until enough confirmations take place to make this unfeasable -- Alice can just send a walk away.

    With a transaction that depends on a whole bunch of Carols in between it becomes the case that Alice, Bob, and all the other Carols have to worry about one of the other Carols double spending in a manner the invalidates Alice's transaction.

    For it to work at all, Alice would have to keep re-connecting and checking that her transaction happened.

    Rippling is a lot more smooth on the Ripple protocol. Naturally this does come at a cost, the ripple consensus process doesn't allow anyone to participate as a contributing validator (you can run the daemon and validate the network, but nobody will care about your work), whereas mining a blockchain is open to all without asking. Meaningful validators on Ripple are listed by other validators as being trusted to not collude with other validators. In order to grant this trust, you really have to know something about the identify of a given validator and not just randomly trust other potential Sybils to not collude.

    Because of this downside, it's going to take awhile for Ripple's validator network to get built out to be fairly distributed and decentralized. And such a network can never be as transaction sending censorship resistant as a blockchain, so blockchain technology is definitely going to be here to stay for direct exchange of tokens and other cool applications.

    The new Stellar design is going to be a little more organic for build-out, but they also acknowledge that it won't have the same censorship resistance properties as proof of work blockchains.
  • b0risb0ris Member Posts: 7
    edited August 2015
    I opened a new thread about Ripple on Ethereum on forum.ripple.com because this thread sucked and received this thoughtful answer:

    "The base to enable Ripple over Ethereum already exists: it's called EtherEx. It's not a 1:1 mapping but EtherEx supports IOUs and distributed order books and matching engine. It would just need to be extended with a trust lines system and have the client do path finding and it could become a decentralized multi-currency payment solution like Ripple.

    The nice thing with Ethereum is that contracts can call other contracts so we can create a payment contract and have it use transparently the EtherEx contracts.

    Since Ethereum is conceived to have DApps run as html+javascript in Mist and communicate with geth using JSON, the client-server interface is very close to what Ripple is using so it would be conceivable to make forks of Ripple Trade and Ripple Charts that can run in Mist.

    A good starting point would be to port ripple-lib to have it format JSON messages in Ethereum format. "
  • b0risb0ris Member Posts: 7
    edited August 2015
    From Twitter:

    Andreas Antonopolous: Ripple could get reimplemented on top of Ethereum in a few lines of code

  • FinancistoFinancisto Member Posts: 8
    Isn't building Ripple on Ethereum something like using horses in order to pull a Ferrari? :*
  • b0risb0ris Member Posts: 7
    edited August 2015
    @Financisto ripple.com uses a blockchain like ledger with consensus algorithms, and they had to invent gas, XRP, that pays for computation and prevents spam and so on. it's centralized tough, and implementing it on ethereum would solve that, and ethereum already has built on gas that pays for computation and prevents spam and so on. for a public ledger version of ripple, an ethereum DApp would be a more decentralized and in my mind a better implementation then ripple.com. non-public-ledger versions like ripplepay and implementations that are closer to Ryan Fugger's vision wouldn't have much use for ethereum, and the public ledger ripple vs non-public ledger ripple is a common debate. i think ripple and trust-lines can be used as a book-keeping system even if its public and don't see why a coin like Bitcoin or some randomly issued coin on ethereum would be better.

    in my mind, and some ethereum-devs seem to have the same vision, ether should be used mostly to pay for computation and not as a monetary system.


    (from https://www.ethereum.org/ether)


    it's impossible to know if and how ether will be used, but if we assume it's used for computation only, then ethereum lacks a monetary system. so, a ripple-credit-system on ethereum could be a useful piece of infrastructure for DAOs

    why do you think its like using horses to pull a ferrari ?
  • FinancistoFinancisto Member Posts: 8
    b0ris said:

    why do you think its like using horses to pull a ferrari ?

    It's just because I don't like that much that centralized structure of ripple as it is today...

    I believe that ethereum has the potential to allow the creation of far more powerful and efficient systems than that Ripple experiment...

    Imo Ripple is already outdated, some kind of "Planned obsolescence".

    But don't get me wrong, despite all that's been said, all experiments with cryptocurrencies are a positive process for our community.
  • FinancistoFinancisto Member Posts: 8
    edited August 2015
    duplicated post erased.
    Post edited by Financisto on
Sign In or Register to comment.