Le Traon, Yves[University of Luxembourg > Faculty of Science, Technology and Communication (FSTC) > Computer Science and Communications Research Unit (CSC) > ; University of Luxembourg > Interdisciplinary Centre for Security, Reliability and Trust (SNT)]
[en] The availability of Distributed Database Management Systems (DDBMS) is related to the probability of being up and running at a given point in time and to the management of failures. One well-known and widely used mechanism to ensure availability is replication, which includes performance impact on maintaining data replicas across the DDBMS’s machine nodes. Benchmarking can be used to measure such impact. In this article, we present a benchmark that evaluates the performance of DDBMS, considering availability through replication, called Under Pressure Benchmark (UPB). The UPB measures performance with diﬀerent degrees of replication upon a high-throughput distributed workload, combined with failures. The UPB methodology increases the evaluation complexity from a stable system scenario to a complex one with diﬀerent load sizes and replicas. We validate our benchmark with three high-throughput in-memory DDBMS: VoltDB, NuoDB and Dbms-X.