Recently Hazelcast rolled out its new version - Hazelcast 3.7. This new version incorporates a lot of performance related improvements that, according to Hazelcast, provide a 30% performance boost compared to its previous versions.
We, at GridGain, consistently spend a significant amount of resources to make sure that the performance of our products show the best results, and hence, are curious to see how close Hazelcast could get to GridGain in terms of performance.
GridGain Professional Edition 1.7.1 and Hazelcast® 3.7 (Open Source) were benchmarked on AWS EC2 using Yardstick framework. The results for throughput, latency, and percentile probes measuring various metrics during benchmark execution were collected.
GridGain vs. Hazelcast Performance Summary
Deadlock Free Transactions: |
61% – 129% higher operations/sec throughput |
Transactional Operations: |
11% – 83% higher operations/sec throughput |
Atomic Operations: |
Put Benchmark: 13% – 24% higher operations/sec throughput Put and Get Benchmark: Depending on configuration, either GridGain or Hazelcast performed slightly better |
SQL Query: |
13% – 27% higher operations/sec throughput |
Benchmarks show that GridGain still outperforms Hazelcast on all atomic and transactional cache operations and SQL-based cache queries except for Atomic Put and Get benchmark, which showed mixed results. Detailed testing results and configurations can be found here.