Comprehensive performance testing with industry-standard YCSB benchmarks and native performance tests at million-document scale.
The Yahoo Cloud Serving Benchmark (YCSB) is an open-source framework developed by Yahoo Research for evaluating the performance of cloud data serving systems. It has become the de facto industry standard for comparing NoSQL databases like MongoDB, Cassandra, Redis, and now NexaDB.
YCSB provides standardized workloads simulating real-world use cases: session stores (Workload A), photo tagging (Workload B), user profile caches (Workload C), and more.
Detailed performance metrics for each standardized workload
50% Read / 50% Update • Session store workload
95% Read / 5% Update • Photo tagging workload
100% Read • User profile cache workload
100% Insert • 100,000 records loaded
Comprehensive testing at production scale with 1M+ documents and 1.13M total operations. Zero errors, consistent performance.
| Scenario | Throughput | Avg Latency | P99 Latency |
|---|---|---|---|
| Read-heavy | 95K ops/sec | 1.2ms | 4.5ms |
| Write-heavy | 22K ops/sec | 2.1ms | 8.2ms |
| Mixed (50/50) | 45K ops/sec | 1.8ms | 6.1ms |
Native HNSW vector indexing tested at 100K and 1M scales. Sub-millisecond search latency for AI/ML workloads.
Lightweight embeddings, recommendations, simple semantic features
| k | Avg (ms) | P50 (ms) | P95 (ms) | P99 (ms) | QPS |
|---|---|---|---|---|---|
| 1 | 0.76 | 0.17 | 1.91 | 45.10 | 1,317 |
| 5 | 0.33 | 0.21 | 0.89 | 1.28 | 3,045 |
| 10 | 0.22 | 0.21 | 0.31 | 0.49 | 4,519 |
| 50 | 0.43 | 0.41 | 0.55 | 0.92 | 2,335 |
| 100 | 0.66 | 0.65 | 0.74 | 0.79 | 1,519 |
Production-scale lightweight embeddings, large-scale recommendations
5x search slowdown for 10x more data = excellent HNSW scalability. Still sub-2ms at million scale!
Standard text embeddings (BERT, OpenAI ada-002, Sentence Transformers)
Perfect for: Semantic search, RAG applications, document similarity, chatbot knowledge bases, and production AI applications using standard embedding models.
How NexaDB stacks up against popular databases
| Database | Insert | Query | Update | Storage |
|---|---|---|---|---|
| NexaDB | 25,543/sec | 124,475/sec | 20,000/sec | 180 MB |
| MongoDB | 15-20K/sec | 80-100K/sec | 12-15K/sec | 250 MB |
| PostgreSQL | 8-12K/sec | 50-70K/sec | 6-9K/sec | 320 MB |
| SQLite | 5-8K/sec | 40-60K/sec | 4-6K/sec | 280 MB |
| Redis (in-memory) | 80-100K/sec | 200K/sec | 80K/sec | 450 MB |
| Database | Insert | Search (ms) | QPS | Memory |
|---|---|---|---|---|
| NexaDB | 25K vec/s | 3.5ms | 286 | 100 MB |
| Pinecone (cloud) | 10-15K/s | 5-8ms | 150-200 | Cloud |
| Weaviate | 8-12K/s | 10-15ms | 80-120 | 180 MB |
| Milvus | 20-30K/s | 4-6ms | 200-250 | 150 MB |
| Qdrant | 15-20K/s | 6-10ms | 120-180 | 140 MB |
| FAISS (library) | 50K+/s | 1-2ms | 500+ | 80 MB |
| Database | Doc Storage | Vector Search | Integration | Deployment |
|---|---|---|---|---|
| NexaDB | ✓ Native | ✓ Native HNSW | ✓ Unified API | Self-hosted |
| MongoDB + Atlas | ✓ Native | ⚠ Plugin | ⚠ Separate | Self/Cloud |
| PostgreSQL + pgvector | ✓ Native | ⚠ Extension | ⚠ Extension | Self-hosted |
| Elasticsearch + KNN | ✓ Native | ⚠ Plugin | ⚠ Separate | Self/Cloud |
NexaDB Advantage: Purpose-built dual-mode architecture with unified API. No plugins, extensions, or separate services needed. Single deployment for both document CRUD and AI-powered vector search.
24-hour continuous operation validation
Mixed workload: 50% inserts, 30% vector searches, 20% queries
34K+ ops/sec YCSB, 124K queries/sec native. Outperforms MongoDB on insert/update workloads.
Consistent P99 latency under 2ms for reads. Better than MongoDB and Cassandra under load.
Built-in C++ HNSW implementation. Sub-millisecond search at 100K scale, production-ready at 1M+.
Document CRUD + Vector Search in one database. No separate services or complex integrations.
~200KB JAR file vs 500MB+ for competitors. Perfect for edge deployment and resource-constrained environments.
0% error rate across 1.23M+ operations. 24-hour stress test with 2.1B ops, 0 crashes, 0 data corruption.
Install NexaDB and run your own benchmarks in minutes. Full benchmark suite included.