论文标题
从基本行为模式中衍生谓词
Derivation of Heard-Of Predicates From Elementary Behavioral Patterns
论文作者
论文摘要
分布式计算有许多模型，并且没有用于考虑全部的统一数学框架。回避此问题的一种方法是从简单的通信和故障模型开始，并将它们用作构建块，以得出该领域研究的复杂模型。因此，我们定义了联合，继任或重复之类的操作，这使得从简单模型构建复杂模型的同时保持表达能力变得更加容易。 为了使这种方法形式化，我们在听觉模型中抽象了复杂的模型和操作。该模型依靠（可能是异步）的弹性；在接收器进入下一轮之前，收到了在给定回合中发送的消息的序列，每轮捕获的消息是在给定回合中发送的。一组称为谓词的序列定义了法律沟通行为 - 也就是说，是一种交流模型。由于所提出的操作对操作模型转换为听觉谓词的行为表现良好，因此我们可以得出构造的谓词的界限，特征和实现。
There are many models of distributed computing, and no unifying mathematical framework for considering them all. One way to sidestep this issue is to start with simple communication and fault models, and use them as building blocks to derive the complex models studied in the field. We thus define operations like union, succession or repetition, which makes it easier to build complex models from simple ones while retaining expressivity. To formalize this approach, we abstract away the complex models and operations in the Heard-Of model. This model relies on (possibly asynchronous) rounds; sequence of digraphs, one for each round, capture which messages sent at a given round are received before the receiver goes to the next round. A set of sequences, called a heard-of predicate,defines the legal communication behaviors -- that is to say, a model of communication. Because the proposed operations behave well with this transformation of operational models into heard-of predicates, we can derive bounds, characterizations, and implementations of the heard-of predicates for the constructions.
