Dynamic data replication: an approach to providing fault-tolerant shared memory clusters

Fast and transparent recovery for continuous availability of cluster-based servers

“…In our previous work [8], we extend the base GeNIMA protocol with fault tolerance mechanisms. These previous techniques build and maintain consistent replicas of system and application state during normal execution.…”

“…In this work, we present FineFRC (Fine-grained Failure Reconfiguration on Clusters), a system that builds on our previous work [8] and automatically reconfigures itself upon failures in the above context of shared memory applications on commodity clusters. FineFRC provides support for transparently recovering all inmemory state at low overhead.…”

“…Our previous work [8] deals with (a) by supporting dynamic replication of all in-memory data to multiple nodes. In this previous work, we report overheads of implementing fault tolerance schemes during normal operation, but not the delays during recovery from a failure.…”

“…Our previous work [8] presents a fault tolerance approach that achieves low overhead by using dynamic replication of application and system state during system operation. Our approach can be summarized as follows.…”

“…We use a previously developed system [8] that achieves (a) by using dynamic replication of data to the memories of multiple nodes of the system during execution. We examine in detail how the runtime system can achieve minimum application interruption, when failures occur.…”