Method of electing an active station in a data-processing system made reliable by replication

Abstract

A method is described for electing an active station from two stations having respective mass memories. According to the method a mass memory is associated with the two stations. An identifier is provided in a respective persistent memory associated with each station having either a present state or an absent state. The identifier associated with one of the two stations is switched from the absent state to the present state when the other of the two stations becomes non-operational. The identifier associated with one of the two stations is switched from the present state to the absent state when the other of the two stations becomes operational. A station that becomes operational is elected as the active station if the identifier associated with the station that becomes operational is in the present state.

Dependencies are inferred from the claim text.