ILP for Micropayments demo

[Guest]

 

[tommytrain]

woah

[Guest]

This is a more polished version of what was demo'd at the ILP W3C meeting.

[tomb]

Wow, wasn't expecting that. Nicely done.

[Zooplankton]

So correct me if I'm wrong but ILP can accept payments in any currency because RCL will have market makers fuelled by XRP infused competitive spreads, doing the conversions? With XRP as the bridge currency?

 

Infrastructure for the Internet of Value.

[cmbartley]

5 minutes ago, Zooplankton said:

So correct me if I'm wrong but ILP can accept payments in any currency because RCL will have market makers fuelled by XRP infused competitive spreads, doing the conversions? With XRP as the bridge currency?

 

Infrastructure for the Internet of Value.

Sometimes. ILP can link any two ledgers in theory so its use goes beyond only bridging ledgers using different currencies. For example, it could link Venmo and PayPal's ledgers so that you could send VenmoUSD and the recipient would receive PayPalUSD. It could in theory do this without a spread as long as there's someone with an account on both ledgers willing to facilitate the spread for free. Presumably, though, the intermediate account would leverage some service fee. 

If there's no direct bridge (i.e. an individual/entity with accounts on both ledgers) then ILP could in theory route the payment through RCL which will have more paths. RCL routing would be more likely for cross-currency payments, for example someone pays in bitcoin and the recipient receives Russian Rubles, than for same-currency payments.

[tulo]

1 hour ago, Zooplankton said:

So correct me if I'm wrong but ILP can accept payments in any currency because RCL will have market makers fuelled by XRP infused competitive spreads, doing the conversions? With XRP as the bridge currency?

 

Infrastructure for the Internet of Value.

Nope, ILP won't necessarily connect with RCL.

It can use any ledger with proper interface to ILP... 

[thinlyspread]

Great video – fun and yet a practical use case shown, which is what ILP needs.

[tommytrain]

I like how it presents the accessible tools for devs to work on or integrate into their own projects. 

Building a dev community around interledger seems like an important step.

[JoelKatz]

You can imagine the ILP universe consists of ledgers, connectors, and things that combine some aspects of both. ILP is the protocol that connectors use to make inter-ledger transactions. XRP can certainly be an intermediary currency for an ILP transaction, but there is no requirement that any particular intermediary, or any intermediary, be used.

For example, imagine if a whole bunch of payment networks including PayPal and Dwolla were ILP enabled. Someone (let's call him Mark) could provide trustless liquidity between PayPal and Dwolla directly, without any intermediary other than them.

Say Alice has money on PayPal that she wants to send $10 to Bob, who accepts only Dwolla. The flow could be as follows:

1) Alice requests a quote from Mark, say it's $11.

2) Alice receives and accepts the quote.

4) Alice uses ILP to tell PayPal to hold $11 of her money, payable to Mark on presentation of a receipt from Dwolla that Bob has been given $10 in Alice's name. The sender need not trust Mark because she knows that PayPal (who holds her money, so she must already trust them) will not release her funds to Mark unless it sees a receipt from Dwolla. (Alice doesn't trust Dwolla, it's Bob's problem if Dwolla "eats" his payment because Bob trusts Dwolla or he wouldn't have asked Alice to pay him there.)

5) Mark doesn't trust Alice or Bob, but must trust both Dwolla (because they hold his money) and PayPal (because he's accepting a payment there). Mark can now pay $10 to Bob, knowing that Dwolla will issue the receipt and PayPal will honor it. When Mark gets the receipt from Dwolla, he gives it to PayPal. PayPal releases Alice's $11 to him.

Now, Bob has received $10 from Alice on Dwolla. Alice has paid $10 on PayPal. The only intermediary was the market maker who accepted the PayPal payment from the sender in exchange for making the Dwolla payment to the recipient.

The core mechanic of ILP is these two operations -- hold money subject to release only in particular ways if particular conditions are met, and release money given proof the relevant condition was met. The core insight of ILP is that those two operations are (almost) all you need a ledger to implement to make cross-ledger payments possible with senders not have to trust recipients, recipients not having to trust senders, and neither senders to recipients having to trust any intermediaries other than the ledgers that hold (or will hold) their money by their choice.

Of course, if there's no better path than XRP for a payment, then ILP can use XRP as a vehicle currency. But it can use anything else too.
 

(I say these two operations are almost all you need because the flow above is not really suitable for large payments. For example, if Alice locks up her $11 forever, then Mark can cause her grief by not doing anything. But if Alice locks up her money with some kind of reasonable timeout, then what happens if there's a failure of a ledger? Mark might make the payment to Bob but not be able to get the receipt from Dwolla or to PayPal in time. For a $10 transaction, he can price that tiny risk into his rates, maybe it will add 0.1%. But for a multi-million dollar transaction, that's not going to work. So ILP has an atomic mode. But this parenthetical is already too long, so I'll explain it some other time.)

[rippleric]

So you're saying there's a chance  How about a video  

How about for crossborder payments @JoelKatz would this still be feasible and cheap enough for a single market maker to facilitate?

Don't you need some type of consensus network to eliminate single points of failure and bridge the rest of the world regardless?

Edited May 4, 2016 by rippleric

[JoelKatz]

6 hours ago, rippleric said:

So you're saying there's a chance  How about a video  

How about for crossborder payments @JoelKatz would this still be feasible and cheap enough for a single market maker to facilitate?

Don't you need some type of consensus network to eliminate single points of failure and bridge the rest of the world regardless?

You wouldn't want me to make a video.

Yes, this is all definitely theoretically doable for cross border payments. I think as a practical matter you'll probably see mostly licensed and regulated exchanges playing in that pond. But you never know, once it's technically possible, who knows what people will do with it. In the early days of the Internet and in the pre-Internet era, only big companies could provide computerized services. Today, anyone who wants to can have a web site.

You do need some kind of consensus network to eliminate single points of failure. But:

1) The consensus network can exist only for that one transaction if you want.

2) The consensus network doesn't have to know that much about the transaction.

3) The consensus network can operate at any speed.

I would expect that you'll see PBFT used as the algorithm and consensus reached in less than a second. The purpose of the network is to provide a resilient way for everyone to agree on whether or not the transaction succeeded.

[cmbartley]

Glad to see that at least two companies are actively working to implement interledger: