スケーリングビットコイン・ワークショップ: Tel Aviv 2019

WIP: Conditional Transfer of Tokens on Top of Bitcoin

Work in progress sessions

Scaling Bitcoin 2019 Tel Aviv "yesod"

https://twitter.com/kanzure/status/1172173183329476609

Conditional transfer of tokens on top of bitcoin

Thomas Eizinger (coblox.tech)

I want to work on trustlessly rebalancing your LN channels with USDT. Let me tell you how lightning channels work. Just kidding ((laughter)). Does anyone remember omnilayer? It was originally called mastercoin. It has tokens on it like USDT. Omnilayer is a separate consensus layer on top of bitcoin, using OP_RETURN to embed extra data into bitcoin transactions where this data is not validated by bitcoin nodes but is validated by omnilayer clients.

Can we use HTLCs with omnilayer simple_send? The answer is yes we can do that. It works just as you would expect it to work. You have an output, and the consensus rules require you to provide the redeem transaction with the HTLC. It provides the preimage to the hash.

How is this useful? We've learned several things while doing this investigation. If you do assets on top of bitcoin that aren't bitcoin, you don't need to include any functionality that gives you conditional transfers. We can define assets on top of bitcoin and still use HTLCs and scripts to conditionally spend those assets. Omnilayer actually, is, it has a lot of commands to do all kinds of things. You can't directly buy tether with bitcoin-- you have to go through a weird omnicoin to buy tether, which I think just is there to make them money. But you can construct a transaction where within one transaction someone else sends bitcoin and you get tokens back. It's an atomic swap on the same chain.

Omni can potentially be used with everything that uses HTLCs such as cross-chain atomic swaps or LN.

Constructing these transactions is non-trivial. You can make a lot of mistakes if you put the wrong address in, the consensus layer may not recognize where it wants you to transfer it. This is where our WIP comes in. We want to avoid one-off implementations. We want to implement the use case of rebalancing your LN channel with USDT and we want to avoid the implementation where all you have to build upon is the blockchain node.

We are working on COMIT, a framework for cryptographic protocols. We're trying to give you primitives that you will need when you build a cryptographic protocol so that you can actually focus on your work. Not just a prototype. Exchanging messages over a robust framework, or negotiating keys, or most cryptographic protocols have timeouts because parties can leave at any point so you need to watch the chain. You need some reusable pieces. You want to hide cryptographic complexity from users.