Foundations of Blockchains (Preview of Lectures 2--7: A Bootcamp on Classical Consensus) скачать в хорошем качестве

Foundations of Blockchains (Preview of Lectures 2--7: A Bootcamp on Classical Consensus)

A lecture series on the science and technology of blockchain protocols and the applications built on top of them, with an emphasis on fundamental principles. Full playlist:    • Foundations of Blockchains   This video: Overview of Lectures 2--7, a Bootcamp on Classical Consensus Accompanying notes: https://timroughgarden.github.io/fob2... Accompanying slides: https://timroughgarden.github.io/fob2... Leave comments/questions below or at   / 1488156743905468421  . Tl;dr: 1. Lectures 2–7 constitute a bootcamp on some of the fundamental work that was done in distributed computing on consensus in the 1980s. This work may be “old,” but it very much informs modern blockchain protocol design. 2. Lecture 2 introduces the synchronous model (with reliable message delivery) and covers the amazingly fault-tolerant Dolev-Strong protocol for state machine replication. 3. Lecture 3 presents our first impossibility result, showing that the guarantees of the Dolev-Strong protocol are provably unachievable without some type of “trusted setup” assumption (like the in-advance distribution of nodes’ public keys). 4. Lectures 4 and 5 introduce the asynchronous model, which makes no assumptions about message delivery (other than eventual delivery), and proves what is arguably the most famous impossibility result in distributed computing: deterministic consensus is impossible in the asynchronous setting. 5. Lecture 6 introduces the partially synchronous model as a sweet spot between the synchronous and asynchronous models, proves that consensus is impossible in this model when more than 33% of the nodes can deviate from the protocol, and compares these results to a famous principle from distributed systems, the CAP theorem. 6. Lecture 7 describes the Tendermint protocol (the basis of the Cosmos and Terra networks, among others) and proves that it achieves optimal fault-tolerance in the partially synchronous setting.