REDUCING OVERHEAD OF SELF-STABILIZING BYZANTINE AGREEMENT PROTOCOLS FOR BLOCKCHAIN USING HTTP/3 PROTOCOL: A PERSPECTIVE VIEW

Today, there is a tendency to reduce the dependence on local computation in favor of cloud computing. However, this inadvertently increases the reliance upon distributed fault-tolerant systems. In a condition that forced to work together, these systems often need to reach an agreement on some state or task, and possibly even in the presence of some misbehaving Byzantine nodes. Although non-trivial, Byzantine Agreement (BA) protocols now exist that are resilient to these types of faults. However, there is still a risk for inconsistencies in the application state in practice, even if a BA protocol is used. A single transient fault may put a node into an illegal state, creating a need for new self-stabilizing BA protocols to recover from illegal states. As self-stabilization often comes with a cost, primarily in the form of communication overhead, a potential lowering of latency - the cost of each message - could significantly impact how fast the protocol behaves overall. Thereby, there is a need for new network protocols such as QUIC, which, among other things, aims to reduce latency. In this paper, we survey current state-of-the-art agreement protocols. Based on previous work, some researchers try to implement pseudocode like QUIC protocol for Ethereum blockchain to have a secure network, resulting in slightly slower performance than the IP-based blockchain. We focus on consensus in the context of blockchain as it has prompted the development and usage of new open-source BA solutions that are related to proof of stake. We also discuss extensions to some of these protocols, specifically the possibility of achieving self-stabilization and the potential integration of the QUIC protocol, such as PoS and PBFT. Finally, further challenges faced in the field and how they might be overcome are discussed.

Distributed Computing Column 81

Overview. In this edition of the column, we have an exciting contribution from Shir Cohen, Idit Keidar, and Oded Naor (Technion), who provide an in-depth perspective on communication-efficient Byzantine Agreement. With the huge popularity of blockchains, the classic area of Byzantine Agreement has seen a surge of interest, and, given the large-scale and widely-distributed deployments of such mechanisms, communication efficiency is a chief concern. This column provides a gentle introduction to the area, by rst the early work in the 80s. This provides the necessary context for the recent exciting work on communication reduction, from the King & Saia algorithm to Algorand. One very useful feature in this column's contribution is the fact that it provides a uni ed view of these results, along with the mathematical background to understand and di erentiate the underlying results. Many thanks to Shir, Idit and Oded for their contribution!

Byzantine Agreement with Less Communication

The development of reliable distributed systems often relies on Byzantine Agreement (BA) [26]. In this problem, a set of correct processes aim to reach a common decision, despite the presence of malicious ones. BA has been around for four decades, yet practical use-cases for it in large-scale systems have emerged only in the last decade. One major application for BA is cryptocurrencies. For example, Bitcoin [30], the rst cryptocurrency, requires a large set of users to agree on the state of the blockchain. Since Bitcoin is a real currency with real value, the need to protect it against Byzantine users is crucial. Following Bitcoin, many other blockchains and FinTech platforms have emerged, e.g., [6,21,30,36]. Consequently, an efficient implementation, in terms of communication, has become one of the main foci of BA solutions.