Ripple Credit System Could Help or Harm Bitcoin

14 January 2013—This Christmas season I bought my very first bitcoins. The idea was to give them as gifts to my family, thus setting the Pecks out on the adventure of figuring out how Bitcoin—the four-year-old decentralized, stateless cryptocurrency—works and why it might be important. I enjoy giving homework as a present.

The adventure, however, turned out to be my own, as I first had to find a way to get some bitcoins. Apparently there are a few ways: You can join the network that creates bitcoins and mint them at a slow, unpredictable pace. You can accept them as payment. You can buy them from a centralized exchange. Or you can find a random stranger online who wants to sell them and meet up with him on the street. I chose the latter, an option that landed me in a coffee shop three days before Christmas, handing over US $100 cash to a man I had never met, whose father had promised to immediately transfer me bitcoins from his home in the Caribbean.

And so he did. But this arrangement is far from perfect, as are all of the alternatives. Each requires that a Bitcoin buyer trust either strangers to follow through on their promises or a centralized exchange to broker the deal. Confidence in Bitcoin’s centralized exchanges is especially wobbly following a series of hacking incidents last year. Bitcoinica and BitFloor, the two hardest hit, are still struggling to pay back their customers. And a larger problem looms: the specter of government regulation. If any country decides to outlaw Bitcoin, bearing down on the exchanges will be the easiest way to do it.

Within a few weeks, a new option will be available: a system called Ripple, which allows individuals to create credit and disburse it to people within a peer-to-peer social network. The project could be used to implement what many Bitcoiners have been asking for—a decentralized currency exchange. But if the more-ambitious parts of the design pan out, the credit created in Ripple could itself become a new form of digital currency and the first formidable competition for Bitcoin.

According to those at the helm of Ripple—including Jed McCaleb, who is best known in the Bitcoin community for developing Mt. Gox, the largest centralized Bitcoin exchange—they will release a public version at the end of January, the first peer-to-peer iteration of the Ripple project. At the moment, Ripple’s developers are testing a beta version on a small group of finance-minded developers, and they were good enough to share access with IEEE Spectrum.

Ripple is actually a concept that predates Bitcoin and was first implemented by Ryan Fugger, a Web developer in Vancouver, B.C., Canada. His idea was to create a way for people to extend credit to strangers through the people they already knew. For example, if I trust Alice, who trusts Bob, I can offer my credit to Bob in an agreement backed by Alice.

When Fugger built Ripplepay in 2005 and 2006, “the goal was to create a system of debt money without artificially imposed scarcity,” he says in an e-mail. What that meant in practice was giving individuals the power to operate as their own banks, with the ability to issue credit.

When you join the Ripple network, you designate primary contacts and establish levels of trust by choosing the amount of credit you are willing to offer each individual. I might feel okay lending up to $1000 to a few close friends whom I know will have a deep obligation to pay me back, while making only $50 available to the man I talk to every morning at the deli. But with Ripple, the credit isn’t only available to my acquaintances. It becomes available to anyone else who knows the people in my network.

Let’s say, for example, that I want to sell a Web application to Bob, who knows my close friend Alice (credit limit $1000). Ripple allows Bob to trade with me using his credit limit with Alice. If she trusts him for $100, that becomes his credit limit with me. I would send Bob the merchandise in exchange for his promise to pay. But instead of owing me, Bob would actually owe Alice, who would in turn owe me. Once Bob satisfied his debt, the Ripple network would destroy the entire chain of IOUs.

“In this way, I can effectively trade with someone I don’t know, whose personal reputation is worth nothing to me, through the people I do know,” says Mike Hearn, a software developer in Zurich who wrote BitcoinJ, a Java implementation of the Bitcoin protocol.

This fall, Hearn gave one of the more technical talks at a Bitcoin conference in London, throwing out a slew of ideas for how to improve the currency. Adapting Ripple was one of them. Specifically, he discussed turning it into a peer-to-peer network and using it to facilitate Bitcoin exchanges between individuals.

“I was suggesting an integration of this idea specifically for the purpose of trading state currencies for bitcoins,” says Hearn.

According to McCaleb, the new version of Ripple will include an exchange platform built right into the network. “For the currency exchange, you create offers, and these offers are stuck in the ledger. Anyone that has a trust path to the currency you are buying can take your offer,” McCaleb explained in an e-mail.

This implementation would solve the problem I had over Christmas. Instead of meeting with a random stranger, I could connect with someone in the Ripple network and buy bitcoins from him or her with an IOU that the person would trust me to settle.

For many who use or want to begin using Bitcoin, this is a welcome development. “If it does achieve that, I’d say that’s pretty big news, because the main weakness of Bitcoin, obviously, is that you have to rely on an exchange,” says Eli Gothill, a Bitcoiner who has played with the new beta version of Ripple and who blogs about digital currencies and payment platforms from his home in London.

Not only will this reduce the risk for individuals who want to trade in bitcoins (or other currencies, for that matter), it will also be much harder to shut down than a centralized exchange, because the upcoming version of Ripple will be the first iteration to run on a peer-to-peer network with open-source software.

“The database that stores all the transactions and relationships is shared across the network. So everybody hosts it. And then every 10 to 20 seconds, there’s a consensus process where everybody’s node is brought up to date,” says Gothill.

So far, Ripple’s founder approves of the new version. “An open, distributed transaction network is a hard problem, and I think Jed’s approach has tons of potential,” says Fugger.

And that potential encompasses much more than a slight improvement for the Bitcoin economy. Embedded in the Ripple concept is the possibility that the network could become so large that its users would start to participate in a closed loop of trade, treating personally issued credit as a unique form of currency. In this scenario, rather than immediately demanding payment on IOUs, Ripple users would pass them on as payment in another transaction. And then IOUs would circulate like currency until they canceled each other out.

This is the bigger idea behind Fugger’s original design. “What Ripple is designed to do is get rid of the need for money to settle debts, because debts can be canceled out against each other. So if I send you an IOU for 10 and you send me an IOU for 10, then we don’t need to settle anything, because our debt will just cancel out,” says Gothill.

With enough small, interconnected trust relationships, or with a handful of universally trusted entities participating in the system, it could work. And that’s why some people who want Bitcoin to succeed consider Ripple to be both a complement and a potential threat. “I think it’s reasonable to say that if it were successful, it would compete with Bitcoin for transactions. People might pay each other in Ripple instead of Bitcoin,” says Gothill.

There are a couple of more incentives to do so. Although the architecture of Ripple will be conceptually similar to that of Bitcoin, by design it will consume much less electricity than Bitcoin and will be able to conduct transactions over the course of seconds rather than minutes.

But even those who have tried the new version are reining in their expectations. The original version of Ripple has been around for a few years now, and it has yet to prove that you can build a mature economy from a network of acquaintances. There are plenty who think it’s a bad idea to make friendship the basis of a credit network.

“Imagine if you had to phone up your friends and ask them to pay up on a $100 balance. I could imagine that could create some awkwardness,” says Gothill. “But then if you can’t rely on friends to build a network, what are you going to rely on?”