NoSQL/NewSQL DB Performance: Updates for June 2013
Following our previous posts, we overviewed the latest performance results achieved by NoSQL/NewSQL vendors.
- MapR accelerates performance of HBase applications by 25x
- VoltDB 3.2: ~1 million transactions per sec with 3 nodes
- NuoDB’s scale-out performance
- TokuMX vs. MongoDB: up to 60x better insertion performance
MapR accelerates performance of HBase applications by 25x
MapR reported that their big data platform backed with Fusion-io’s Flash memory technology accelerates performance of read-intensive HBase applications by 25x. The solution that combines MapR M7 with Fusion ioMemory eliminates disk storage bottlenecks that used to limit HBase adoption for high-performance systems.
VoltDB 3.2: ~1 million transactions per sec with 3 nodes
VoltDB have revealed performance results for VoltDB 3.2. The team measured transaction per second vs. latency on a sample application. The post on their official blog demonstrates how to set up a cluster that can process ~1 million transactions per second with just three VoltDB nodes.
NuoDB’s scale-out performance
NuoDB have used DBT-2, an implementation of the TPC-C benchmark, to test their NewSQL system. The results published on the company’s official blog compare the number of new order transactions per minute against the number of warehouses based on NuoDB Transaction Engines. Have a look at the figures to evaluate NuoDB’s linear scalability.
TokuMX vs. MongoDB: up to 60x better insertion performance
Tokutek, database performance accelerator, have benchmarked how their TokuMX, “performance-enhancing software,” increases the speed of MongoDB. The test demonstrates up to 60x improvement in insertion performance due to TokuMX’s patented Fractal Tree™ Indexing. Previously, Tokutek announced 268x query performance increase and 532x faster multikey index insertion for MongoDB.
These were the most notable performance updates from NoSQL and NewSQL vendors in June 2013. Follow our blog for more.No Comments