Shared-Nothing Approaches

Abstract

This chapter analyses high availability approaches based on the shared-nothing hardware architecture. A main advantage of the shared-nothing approach is that two shared-nothing sites with two replicas of a replication unit can provide a high degree of failure isolation with respect to software failures and environmental disasters, e.g. fires, floods, avalanches, and land slides. The sites can also be managed independently, which provides a high degree of isolation from operator errors.