RedisGears
summary
RedisGears is a Redis module that runs Python map/reduce pipelines inside Redis using embedded CPython sub-interpreters and a custom allocator to limit per-executor memory. It supports readers (keys, values, streams), map/filter/flatmap/countby, and cluster-aware shuffling so a genre count over ~30k movie records runs unchanged on a 3-shard cluster. It also supports a streaming API for event-triggered pipelines and integrates with RedisAI for in-Redis image classification; caveats include shared global state and some serialization limits for native objects.
AI-written summary · about
chapters (12)
- 0:00Introduction
- 5:20What is RedisGears
- 8:30How operations work
- 12:04Use cases
- 23:18Working with clusters
- 27:23Data streaming
- 31:04Python pros and cons
- 32:28Memory management
- 34:17Cluster data serialization
- 35:45RedisAI integration
- 37:11Practical classification example
- 39:00Q&A
related talks
59mUsing Redis beyond caching
51mX-Raying Your Redis App: SDK Tuning with a Performance Engineer’s Eye
25mpersistent collections as a service
24mWhen Real Time Meets Big Data & Analytics
36mBuild a Cloud Native Asynchronous Messaging System For Scale With Redis
33mIntroduction to Redis
56mRedis Application Patterns in Rails
34mTaking deep learning to production with RedisAI