A Replication Model for Trading Data Integrity against Availability

Abstract

Higher availability and better performance of data-centric applications can be achieved by replication of objects or data items. If data integrity, a correctness criterion for such systems, needs to be maintained even during degraded situations (node or link failures) the system soon becomes (partially) unavailable. However, some applications exist (e.g., in control engineering) where data integrity can be relaxed for higher availability during degraded situations. Traditional replication models do not support the balancing of these two properties. In this paper, we present a novel replication model that (i) allows replicas to diverge if data integrity can be temporarily relaxed and (ii) re-establishes both replica consistency and data integrity during repair time