. . . . "Abstract: \"There has been a wealth of research on reliable message transmission for distributed publish-subscribe systems. However, little research has been done on building fault-tolerance for these systems. Although distributed publish-subscribe systems are intrinsically fault-tolerant, as publishers can be easily replicated on multiple nodes, an event is not guaranteed to be delivered when a node crash happens to one of the subscribers: there is no way for the publisher to distinguish a node crash from a node departure. This research has two goals: first, to add fault-tolerance on existing distributed publish-subscribe systems; and, second, to analyze the tradeoff between scalability and the imposed fault-tolerance based on our implementation of the system.\""@en . . "Fault tolerance for distributed publish-subscribe systems"@en . . . . . .