Problem
Different NoSQL databases make different tradeoffs for read latency, insert latency, and throughput. The project examined those tradeoffs with a repeatable test strategy.
Storage Systems | YCSB | OpenStack
Benchmarking MongoDB and HBase with YCSB.
A data storage and management project comparing NoSQL database behavior under repeatable Yahoo Cloud Serving Benchmark workloads.
2
NoSQL databases
YCSB
benchmarking tool
D
workload focus
Approach
MongoDB and HBase were installed on OpenStack instances and exercised through YCSB workloads, with the results captured across throughput and latency dimensions.
Measurement
The analysis compared overall operations per second, average insert latency, average read latency, and the effect of total operation count under Workload D.
Outcome
The project produced an evidence-based comparison and a test plan that connected database architecture choices to measurable performance behavior.
Why it still matters
Benchmarking work builds the same instincts needed in production platform engineering: define the workload, control the test environment, measure the right signals, and avoid architecture decisions based only on vendor preference.
Open legacy project archive