Scalability and Composability in Distributed Ledgers

点击翻译按钮获取中文摘要

Blockchain is the next frontier of distributed computing, and its rapid ascent from white papers to top-tier publications and multi-million dollar startups has caught the attention of both academics and practitioners. Much of the discussion surrounding blockchains, and more generally distributed ledgers, revolves around two fundamental questions: (i) will they scale to meet current and emerging demands, and (ii) can we trust their correctness? These questions will be answered through a multi-faceted research program that will engage both empirical and theoretical methodologies. NSERC-funded research activities will address the following short-term objectives: Objective 1: Scalability through composability. Scalability has been recognized as a major weakness of distributed ledgers, leading to a search for designs that compose multiple independent services by combining their outputs carefully to obtain a global ordering over transactions. This objective will focus on developing an understanding of how such compositions can be achieved correctly, and how they affect the inherent trade-off between throughput and latency. Objective 2: Streamlining execution in the absence of failures. Blockchain protocols are complex as they must handle both concurrency and failures, and can be simplified substantially in absence of failures. This objective will investigate performance optimizations that speed up blockchains in failure-free operation while maintaining correctness even under adverse conditions. Modular designs will be developed that can be combined with a variety of existing blockchain protocols. Objective 3: Understanding consistency-latency trade-offs. A blockchain protocol must validate transactions to prevent double-spending and ensure data consistency. This objective will develop an understanding of how much latency is required to detect double-spending within a given probability threshold, which can be viewed as a consistency-latency trade-off similar to those encountered in NoSQL storage systems. Objective 4: Foundations and applications. This objective will examine the theoretical foundations of distributed ledgers, and gain a deeper empirical understanding of their performance trade-offs. Research outcomes may include rigorous definitions of correctness properties, proofs of theoretical complexity bounds, and empirical performance evaluations. The proposed research program will elevate Canada to the forefront of research in blockchain technology by generating concrete software prototypes as well as scientific publications describing protocols, proofs of correctness, and performance comparisons. Trainees engaged in NSERC-funded activities will flourish in an experiential learning environment where they will work hands-on with state-of-the-art cloud infrastructure and interact with industry partners, while developing technical skills vital to Canada's increasingly data-driven economy.