Foundations of bitcoin-compatible scalability protocols

Abstract

Permissionless blockchains allow mutually untrusted users to transfer money in a decentralized way. Unfortunately, these blockchains face a scalability problem, which means they are technically limited to processing only a relatively small number of transactions compared to traditional, centralized systems. Payment Channel Networks (PCNs) are among the most prominent solutions to mitigate these scalability issues.The basic idea of PCNs is to outsource transactions to so-called payment channels between two users and then link these channels to form a network where any two users connected by a path of channels can perform transactions. The advantage is that only the transactions for opening and closing these channels need to go on the blockchain, while any other transaction happens outside of the blockchain, thus increasing the overall transaction throughput. Several different PCN protocols exist and are used in practice (e.g., the Lightning Network with a value of approximately 150M USD). However, even these have their sets of issues. In this thesis, we investigate existing PCN protocols and identify issues in terms of security, privacy, efficiency, and limited functionality. Simultaneously, we introduce new protocols that overcome these issues and improve the state of the art. We focus on Bitcoin-compatible solutions since Bitcoin is not only the largest cryptocurrency in terms of market capitalization but also has a limited set of scripting capabilities, thus making our protocols compatible with a large number of other cryptocurrencies as well. We also conduct a rigorous formal security analysis of our protocols. More concretely, this thesis makes the following contributions. First, we introduce Sleepy Channels, enabling secure payment channels even when users are not continuously online. This is a significant shift from existing constructions where going offline puts users' funds at risk. Further, we generalize the notion of payment channels (Generalized Channels) and make them support any application that the underlying blockchain supports rather than only payments. Second, we introduce a new construction (Blitz) that achieves secure payments across a path of multiple channels in PCNs while reducing the number of interactions for each intermediary to a single one (from two or more) and takes only constant time instead of linear in the path length. We also provide the first secure construction for atomically updating multiple channels that are not on a path (Thora). Finally, we provide the first Bitcoin-Compatible Virtual Channel construction. These virtual channels allow two users to open a direct channel via one intermediary without putting an opening or closing transaction on the blockchain. We further analyze other existing virtual channel constructions, identify a novel attack, and introduce secure and efficient virtual channels over multiple intermediaries (Donner).These contributions interplay seamlessly. Collectively, they offer a versatile, ad-hoc solution connecting any two users securely without an on-chain footprint for applications that go beyond payments. Thus, this thesis aims to reshape the understanding of PCNs and give more general and efficient solutions to the scalability problem.