作者: John Klein , Ian Gorton , Neil Ernst , Patrick Donohoe , Kim Pham
关键词:
摘要: The choice of a particular NoSQL database imposes specific distributed software architecture and data model, is major determinant the overall system throughput. performance in turn strongly influenced by how well model query capabilities fit application use cases, so system-specific testing characterization required. This paper presents method results study that selected among three databases for large, healthcare organization. While considered consistency, availability, partition tolerance (CAP) tradeoffs, other quality attributes influence selection decision, this reports on evaluation results. In our testing, typical workload configuration produced throughput varied from 225 to 3200 operations per second between products, while read operation latency factor 5 write 4 (with highest product delivering latency). We also found achieving strong consistency reduced 10-25% compared eventual consistency.