Highly adaptable dynamic quorum schemes for managing replicated data

Dynamic data replication schemes perform superior compared to static schemes in terms of operation availabilities. However, most of them are bounded wrt. the replication degree, i.e. "classic" dynamics varies the degree in the range from one to n replicas not allowing to exceed this upper bound at run-time. In this paper, we present a new framework for adaptable dynamic replication schemes that is able to overcome this limitation by means of creating and deleting an arbitrary number of replicas at run-time. The new framework conceptually extends the dynamic general structured voting framework. The underlying concepts are presented and an example scenario is provided that compares the adaptable approach to the dynamic grid protocol via simulation. It is shown that adaptiveness increases operation availability while providing cost-efficiency.