RedisJSON Performance Benchmark Compared with MongoDB and ElasticSearch

Overview

Recent official reports demonstrate that RedisJSON (combined with RediSearch) dramatically outperforms other NoSQL solutions. In isolated write tests RedisJSON is 5.4× faster than MongoDB and over 200× faster than ElasticSearch; in isolated reads it is 12.7× faster than MongoDB and over 500× faster than ElasticSearch.

Query Engine

The development of reresearch and RedisJSON emphasizes performance stability across versions. Version 2.2 improves load and query speed by 1.7× over 2.0, while also enhancing throughput and data‑load latency.

Loading Optimization

Benchmark results from the New York City taxi workload show consistent performance gains in each new reresearch release.

Full‑text Search Optimization

Indexing 5.9 million Wikipedia abstracts and executing full‑text queries demonstrates that moving from v2.0 to v2.2 yields substantial improvements in write, read, and search latency, thereby increasing overall Search and JSON throughput.

Comparison with Other Frameworks

Using the YCSB benchmark, RedisJSON, MongoDB, and ElasticSearch were evaluated on document storage, local/cloud availability, professional support, scalability, and performance. The test environment consisted of four m5d.8xlarge VMs (one client, three database servers) on AWS.

MongoDB 5.0.3 : three‑member replica set with text index for string search.

ElasticSearch 7.15 : 15 shards, query cache enabled, RAID‑0 on two NVMe SSDs.

RedisJSON* : RediSearch 2.2 + RedisJSON 2.0 on an OSS Redis Cluster v6.2.6 with 27 shards across three nodes.

100% Write Benchmark

RedisJSON’s ingestion speed is 8.8× faster than ElasticSearch and 1.8× faster than MongoDB, maintaining sub‑millisecond latency for 99% of requests. It uniquely updates its index on every write, unlike ElasticSearch’s near‑real‑time (NRT) approach.

100% Read Benchmark

RedisJSON reads 15.8× more operations than ElasticSearch and 2.8× more than MongoDB, while keeping latency in the sub‑millisecond range.

Mixed Read/Write/Search Benchmark

In realistic mixed workloads (65% search, 35% read, 10% write), RedisJSON maintains stable throughput and latency, while ElasticSearch’s performance degrades sharply as write ratio increases. RedisJSON achieves up to 50.8× higher ops/sec than MongoDB and 7× higher than ElasticSearch, with latency reductions of up to 91× versus MongoDB and 23.7× versus ElasticSearch.

Full Latency Analysis

Across all percentiles, RedisJSON consistently delivers sub‑millisecond latency. In p99, RedisJSON records 0.23 ms, compared to MongoDB’s 5.01 ms and ElasticSearch’s 10.49 ms. Write latency remains sub‑millisecond for RedisJSON and MongoDB, while ElasticSearch exhibits >10 ms tail latency.

How to Get Started

Developers can create a free Redis cloud database in any region or use a RedisJSON Docker container. Updated documentation and client drivers for Node.js (node‑redis), Java (Jedis), .NET (NRedisJSON/NRediSearch), and Python (redis‑py) are available to accelerate adoption.