This article is on reliable broadcast in asynchronous signature-free $n$-process message-passing systems in which up to \tb processes are Byzantine and where, at the network level, for each message broadcast by a correct process, an adversary can prevent up to \tm processes from receiving it (the integer \tm defines the power of the message adversary). So, differently from previous works, this article considers that not only computing entities can be faulty (Byzantine processes), but also that the network can lose messages. To this end the article first introduces a new basic communication abstraction denoted \kl-cast and studies its properties in the previous bi-dimensional adversary context. Then, the article deconstructs existing Byzantine-tolerant asynchronous broadcast algorithms and, with the help of the \kl-cast communication abstraction, reconstructs versions of them that tolerate both Byzantine processes and message adversaries. Interestingly, some of these reconstructed algorithms are also more efficient than the Byzantine-tolerant-only algorithms from which they originate. The article also shows that the condition $n>3\tb+2\tm$ is necessary and sufficient (with signatures) to design such reliable broadcast algorithms.
翻译:此篇文章是在无签名的零用美元处理电文传递系统中进行可靠广播的, 在这种系统中, 直至 Byzantine 是 Byzantine, 在网络一级, 每条电文通过正确程序广播, 对手可以阻止到\ tm 的电文接收( 整数\ tm 定义了电文对手的力量 ) 。 因此, 与先前的作品不同, 此篇文章认为, 计算机实体不仅可能有错误( 双占庭进程), 网络也可能丢失信息。 为此, 文章首先引入了一个新的基本通信抽象记号\ kl 播送, 在网络一级, 每条电文通过正确程序广播, 对手可以防止到 \ tm ( 整数/ t) 的电文解剖过程, 在\ kl- 播送的电文摘要帮助下, 重建它们既能容忍 Byzantine 进程又能接收对方信息。 有意思的是, 这些重的算法也比 Byzantine- disqals) 更有效率, 也显示了这些设计所需的可靠 。