Algoritmo Concurrente por Conjuntos de Pilas con Multiplicidad: SetStackLogic
share
This article aims to describe and explain the theoretical foundations of concurrent and set concurrent algorithms, considering an asynchronous shared memory system where any number of processes can crash. Verification of concurrent algorithms is often described in terms of their progress condition, which guarantees that eventually something good will happen, also called the security of the algorithms, and correctness, which guarantees that nothing bad will happen, also called liveliness. of the algorithms. The meaning of correctness of a concurrent algorithm is explained in detail, focusing on linearizability, and a generalization is addressed, concurrency by sets; which is much more recent and less well known. The SetStackLogic algorithm is shown, which is a set-concurrent algorithm and is also an implementation of a stack with multiplicity. The properties of the algorithm SetStackLogic are demonstrated in a formal and detailed way, in order to present a rigorous scheme in the formalization of this type of algorithm; same that could be used for other algorithms. In addition, the operation of the algorithm is explained through scenario examples that illustrate its dynamics in some possible executions.
READ FULL TEXT