Scalable solutions for next-generation distributed applications

在区块链技术的支持下，下一代分布式应用程序使任何人都可以参与金融等活动，而无需来自政府或可信组织的任何许可、帐户或标识。重要的是，这些应用不需要依赖任何中介/受信任的第三方，因为它们可以依赖区块链技术提供的信任。然而，这项革命性的技术必须先扩大规模，然后才能广泛采用。例如，比特币的区块链每秒可以处理多达10笔交易，而Visa等主流支付系统每秒可以处理数万笔交易。这项研究的目标是通过设计可扩展的区块链解决方案，为下一代分布式应用的主流采用铺平道路。我们通过两个主要主题来实现这一目标。在主题1中，我们重点关注区块链系统的网络层。这是负责传播块和事务的层，因此对可伸缩性至关重要。此外，该层对于区块链的安全和隐私至关重要，并施加了独特的安全要求，如抵抗拒绝服务和拓扑隐藏。这使得为区块链设计快速信息传播方法特别具有挑战性，并与我在过去15年一直致力于的传统网络不同。在本研究的主题2中，我们将探索所谓的第二层解决方案。这些解决方案允许各方在缓慢的区块链之外进行交易，同时依赖区块链上的智能合同来解决任何潜在的纠纷。例如，支付渠道允许Alice和Bob在不涉及区块链的情况下在彼此之间转移支付。创建支付渠道的智能合同保证这些离链转账是可强制执行的；例如，如果Bob以过时的状态关闭渠道，Alice将能够通过向智能合同呈现渠道的最新状态来进行争议。在过去的四年中，我们设计了几种新的方法来提高第二层解决方案的效率和安全性。然而，在这个新领域，还有许多剩余的挑战。例如，参与第二层解决方案的各方通常被要求保持在线，并观察区块链是否存在可能的欺诈。决定下线的一方应该聘请第三方，即瞭望塔，来监视区块链并保护她的利益。然而，设计一份瞭望塔协议是具有挑战性的，因为在数字世界中，一方和受雇的瞭望塔之间往往没有信任。这些挑战的解决方案很快就会付诸实践，进而推动下一代分布式应用程序的主流采用向前迈进一步。

Powered by the blockchain technology, next-generation distributed applications enable anyone to participate in activities such as finance without any permission, accounts, or identifiers from a government or trusted organizations. Importantly, these applications do not need to rely on any intermediary/trusted third-party as they can rely on the trust offered by the blockchain technology. This revolutionary technology, however, must scale before it can be widely adopted. Bitcoin's blockchain, for instance, can handle up to about ten transactions per second, whereas mainstream payment systems, such as Visa, can handle tens of thousands of transactions per second. The goal of this research is to pave the way to mainstream adoption of next-generation distributed applications by designing scalable solutions for blockchain. We approach this goal in two major themes. In Theme 1, we focus on the network layer of blockchain systems. This is the layer that is responsible for propagating blocks and transactions, hence is critical for scalability. Furthermore, the layer is critical for the security and privacy of the blockchain, and imposes unique security requirements such as denial of service resistance and topology hiding. This makes designing fast information propagation methods for blockchain particularly challenging and different from traditional networks, which I have been working on over the past 15 years. In Theme 2 of this research, we will explore the so-called Layer-two solutions. These solutions allow parties to conduct transactions outside the slow blockchain, while relying on a smart contract on the blockchain to resolve any potential dispute. For instance, a payment channel allows Alice and Bob to transfer payments between themselves without involving the blockchain. The smart contract that created the payment channel guarantees that these off-chain transfers are enforceable; if Bob, for example, closes the channel with an outdated state, Alice will be able to dispute by presenting the latest state of the channel to the smart contract. Over the past four years, we have designed several new methods to improve the efficiency and security of Layer-two solutions. There are, however, many remaining challenges in this new field. For instance, parties involved in Layer-two solutions are often required to remain online and observe the blockchain for possible frauds. A party who decides to go offline should employ a third-party, referred to as watchtower, to watch the blockchain and protect her interest. Designing a watchtower protocol, however, is challenging, because in a digital world, there is often no trust between a party and the employed watchtower. A solution to such challenges quickly finds its way to practice, and in turn pushes the mainstream adoption of next-generation distributed applications one step forward.