Beyond Logical Tree Structures for Data Replication Schemes beyond Logical Tree Structures for Data Replication Schemes

General Structured Voting is a general framework for a wide range of data replication and distributed mutual exclusion schemes. It allows an easy tailoring for speciic objects and environments in a homogeneous manner. In this paper we investigate the characteristics of a so-called voting structure used by General Structured Voting to derive sets of cooperating nodes. Such a set of cooperating nodes guarantees a correct operation execution on a replicated data object. A voting structure can be represented as a hierarchical, acyclic graph consisting of physical and virtual nodes. Whereas former concepts known form the literature only take advantage of { at most { hierarchical tree structures, we show that the use of a voting structure is clearly more exible, i.e. it allows the construction of a wider variety of sets of cooperating nodes. This exibility has a direct impact on the number and kind of replication schemes which can be modelled. Additionally, we present a hierarchy of concepts relating Weighted voting, Coteries, and General Structured Voting using both, tree-shaped voting structures and the more general hierarchical, acyclic voting structures to each other.