The Raiden Network - Lefteris Karapetsas

so uh we'll make a lightning talk about the riding Network um which uh is um something that um the brainbot team is working on uh of which I'm a member so essentially what um the r network is it's it's trying to to address some of the main ethereum problems which is well not only ethereum main blockchain problems which are for example uh main one is scalability another one is finality of transactions and of course um privacy of transactions the way with which we are trying to do this um is to use the solution of offchain transfers essentially we are using um bilateral transfers between many different nodes that do not need to be on the blockchain and the blockchain is only a Ledger on which the final computations are stored um so the writing network is essentially um a network that can transfer any assets uh it's uh ispired and very similar to the Bitcoin lightning Network it's developed by by brainbot a um uh two developers of which work in from this office uh we are a work in progress and we are working towards a minimum viable product at the moment what are the the features that rting has essentially you can transfer any erc20 compatible token for example I don't know this x token or uh maker die um you can have instant asset transfers and micro payments because essentially nothing needs to go in the blockchain unless there is um uh um a challeng and uh it can also scale to even 1 million transactions per second because um it does not need to um be included in the blockchain unless there is um a settlement and transactions are also confidential how does um such a network work essentially we utilize something called payment channels um a payment tannel in ethereum would be a smart contract onto which a security deposit is given it can be settled closed but um by any of the two parties um and um there can be only a challenge period into which um someone can challenge and get um the deposit back if if there's a problem it looks something like that for example there would be um there will be a contract with a deposit given by both um uh participants it would know the receiver and the initiator address and um the the the participants essentially sign transactions but do not put them in the blockchain but they only keep um how much of an asset has been spent they know it U they both know how much they have spent um a bidirectional channel uh so channels are B directional and it looks like this essentially there is two deposits inside the the contract um and each participant has to track both the balances so Alice and Bob here both track Alice and Bob's respective balances and the um they can spend as much as is the the deposit inside the contract and we have um a net distributable of of um the balance of Alice for example here in the in the graph um essentially in order to scale this there can be many um channels that are connected um so in order to scale this into a full uh Network we have multiple channels that are connected and they mediate transfers between one another um this is what we call a mediated transfer and essentially if Alice wants to send um something to um uh D May say maybe um it can use Bob and Charlie but they would need to have um something called a h loock to sign to um sign a transaction with so that nobody can um um look at it uh unless it has reached the end of of of the of the Channel Once the transaction goes to um to David then he asks Alice for the secet and it goes backwards and uh the transaction um is uh finalized in immediate transfer it's not a small feet depending on um on the distance that's something that we have not implemented yet but yeah that's just how then the idea is supposed to work um so imagine a network of many such with uh Roots being found between any of the two participants uh as long as the transfers cancel out so the deposit is not um so the the the sum of the transfers is not more than the deposit the channel can be long lived so nothing would need to go in the blockchain uh which would result in having a very low frequency of settlements uh that would uh make everything much cheaper because there would be no gas cost since nothing goes in the blockchain um fast for the same reason and there would still be no counterparty risk because the the transactions actually do exist and they are cryptographically proven and each one depends on the previous one they just are not put in the blockchain unless there's a problem um there are many applications for such a thing one example would be a decentralized exchange for example if you want to trade two assets you would need to have a channel open for both of those assets um the initiator would would um broadcast the intent for example he would have a bat and an ask for for asset a and asset B then interested parties would uh send an exchange request and the initiator would accept and the way that this would work is that um we would essentially have two Medi transfers with the same HL so that once the secret is revealed and if and only if it is revealed then either both of the transactions would execute or or if um there is a problem none of them would be executed um as I said in the beginning this is inspired by the lighting Network now if we compare both of them um in ethereum uh R has an easier implementation thanks to the virtual machine um if there is an un comparative participant the channel is not locked we just find another um uh uh route um we can add more funds to a Smart contract uh so the channel can get much more um funds if needed if the deposit is not enough we can use it for any compatible token with the lc20 um uh token proposal and it's interruptable with smart contracts in general we can use oracles and anything else you you may need um the current status is that we have demonstrated 500 transfers per CPU code per second it's a work in progress and um we don't need anything extra added to ethereum at the moment we just need to implement it that we don't wait for Metropolis or anything like that um our Milestone is pro of concept zero we we have demonstrated with the test net a transfer from uh Copenhagen to Mumbai mediated by a node in Florianopolis Brazil um and this happened around August I think um we're heading towards the minimum viable product which is the big milestone that we want to come uh to that would come out around the beginning of 2017 we would like to um if you like Python and you like ethereum we would really appreciate the help the it's in GitHub it's riding Network SL riding and um yeah that's that's all do you have any questions [Applause] on