A new open-source caching software, Pogocache, recently reached 1.0 general availability, focusing on low latency and CPU efficiency. Pogocache supports multiple popular protocols while claiming better throughput and lower latency than other open-source caching alternatives.
Started by Josh Baker, creator of the geospatial database and geofencing server Tile38 and the TG geospatial library, Pogocache is written in C and runs as a server-based program. Pogocache supports Memcache, Valkey/Redis, HTTP, and PostgreSQL wire protocols, allowing the use of system tools such as curl and psql, as well as third-party client libraries. According to the author, Pogocache is a fast caching solution optimized for extremely low contention and good memory locality. Baker writes:
Pogocache is faster than Memcache, Valkey, Redis, Dragonfly, and Garnet. It has the lowest latency per request, providing the quickest response times. It’s optimized to scale from one to many cores, giving you the best single-threaded and multithreaded performance.
As with most caching solutions, the target use cases include session storage, database query caching, real-time data feeds, and API acceleration.
Performance of Pogocache against other caching software running 8 threads on an AWS c8g.8xlarge. Source: Pogocache GitHub
The project includes a section with design details that highlights a sharded hashmap inspired by the author’s shardmap project and significant improvements in thread handling. Pogocache stores entries (key-value pairs) in a sharded hash map that has a high fan-out, often with 4096 shards. The number of shards is automatically configured at startup, but can be changed by the user, with each shard having its own independent hashmap. That hashmap uses open addressing with Robin Hood hashing. Baker adds:
Instead of running Pogocache as a server-based program, the self-contained pogocache.c file can be compiled into existing software, bypassing the network and directly accessing the cache programmatically. Running embedded provides raw speed, with over 100M ops per second.
Pogocache determines the number of threads to use for the program’s lifetime at startup, based on the machine’s core count. Simon Willison, co-founder of Lanyrd and co-creator of the Django web framework, comments:
Performance aside, the most interesting thing about Pogocache is the server interface it provides: it emulates the APIs for Redis and Memcached, provides a simple HTTP API and lets you talk to it over the PostgreSQL wire protocol as well!
On Hacker News, most feedback is positive, with some asking for more design details to support the performance claims. According to the documentation, the protocols are automatically detected, eliminating the need to manage multiple ports.
psql -h localhost -p 9401
=> SET first Tom;
=> SET last Anderson;
=> SET age 37;
$ curl http://localhost:9401/last
Anderson
Source: Simon Willison’s Weblog
On his blog, Cezar Henrique da Costa e Souza warns:
While the initial release and performance claims are undeniably exciting, the true test for Pogocache lies ahead. How will these benchmarks hold up under rigorous, independent scrutiny across varied production workloads? How quickly will the developer community embrace and contribute to the project?
Pogocache’s roadmap includes building domain-specific integrations for web, SQL, prompts, and geo, adding direct access over shared memory and IPC, and providing enterprise tooling such as distributed routing and failovers.
The recently released version 1.1 adds automatic background sweeps of expired keys. Pogocache is open-source software licensed under the AGPLv3 license.
