Same data — a shorter window to bring a new replica up.
Redis protocol compatible · A memory-efficient drop-in
Run the same Redis
on half the memory
Engula is a Redis-protocol-compatible drop-in kernel. No client code changes, works with your existing tooling, no vendor lock-in.
Prove the gains on your own data, and roll back at any time if it does not meet your expectations.
99.40% match on the official test suite
We run the official Redis 7.2.11 test suite directly: 4300+ subtests match, with every difference and its impact disclosed item by item.
~50% less memory
Typical workloads save 40%–75% memory, depending on data structures, value sizes and key count.
RDB path measured in the open
Fixed release baseline: single-point BGSAVE 13.7x, RDB LOAD up to 6.1x (all vs Redis 7.2.11). Peak SAVE and other absolute values behind the footnote; full matrix on the live proof page.
Live proof · LIVE
Live verification dashboard
Every 1–2 minutes we automatically run a round of compatibility, performance, high-availability, RDB and other test suites. The stdout stream is fully public, line by line.
All results are publicly traceable at /proof— you are welcome to verify them.
Replication · measured
Faster replication, steadier reconnect
Headline results from two measured scenarios: full-sync and reconnect storm.
Under a reconnect storm and the same BGSAVE, far less extra memory on the primary.
Getting started
Zero-change verification in three steps
From download to quantified gains in as little as 5 minutes.
01
Download and start
The binary works out of the box and your existing clients connect directly. Supports CentOS 7+ / Ubuntu 18.04+ / mainstream Linux distros, free forever.
02
Import a snapshot to assess
ValueSight parses your RDB and produces a per-key memory comparison in 5 minutes, so you can quantify the gains before migrating.
03
Shadow-verify, then switch
Mirror production traffic to compare consistency, switch the primary only once confirmed, and roll back any time.
Measured in production
Quantifiable production gains
Memory-savings data verified in real cross-industry deployments. Every case went live only after shadow-mode verification.
Production case
Real-time fraud detection ·Fintech
74%
Memory saved
Billions of keys, average value < 256 bytes, a high-frequency read/write mixed workload. Switched the primary after 2 weeks of shadow-mode verification.
Production case
Recommendation-engine cache ·Video & entertainment
43%
Memory saved
Hundreds of instances, mostly Hash and ZSet. The 7.7 TB of memory saved translates directly into deferrable capacity purchases.
Production case
Message-stream system ·E-commerce
67%
Memory saved
BGSAVE dropped from 45s to 8s and RDB recovery from 90s to 12s. Shorter maintenance windows mean fewer business interruptions.
Reproduce it on your own data
Assess your RDB with ValueSight
Upload or locally parse your RDB snapshot and get a per-key Engula-vs-Redis memory comparison within 5 minutes. No need to deploy Engula or shift any traffic.
- 01Run ValueSight with Docker
- 02Point it at your dump.rdb
- 03View the per-key / per-type comparison report
Companion tools
A full set of open tools — each reproducible on your own data.
Start with ValueSight to quantify the gains, then use the other tools to verify compatibility, performance and recovery one by one. Each runs with a single Docker command.
$ docker run -it --rm -v "$(pwd)/dump.rdb:/tmp/dump.rdb" -v "$(pwd)/analysis_output:/engula/analysis_output" registry.cn-guangzhou.aliyuncs.com/montplex/engula-valuesightMEMORY-SAVINGS ASSESSMENT — IMPORT AN RDB AND GET A PER-KEY REPORT IN 5 MINUTES
Engula ValueSight
Import your RDB snapshot and generate an Engula-vs-Redis per-key memory comparison report in 5 minutes, so you can quantify the gains before migrating.
- ·Parsed locally, no data uploaded
- ·Per-key / per-type distribution
- ·CSV / JSON export
Get started now
Verify first, then decide.
Engula is free to download, forever. Complete one round of verification on your own real data: memory gains, compatibility and recovery behavior — all the evidence a migration decision needs, in one pass.
Prefer to review the evidence first? See live proof →