django-dynamodb-sessions is ready!
After much tweaking, hand-wringing, and thumping around on staging, django-dynamodb-sessions is ready. The package provides a super-high-scalability session store for those hosting on Amazon’s EC2.
This may be useful to you if…
You’re wanting to trim your DB size, or reduce the read/write thrashing that happens with DB-backed sessions. You also don’t feel like maintaining Redis, or something similar.
With the introduction of Amazon’s DynamoDB, we now have a very fast, massively scalable data store that is somebody else’s problem to secure, update, and keep running. I’ve found the response times to be great, especially when combined with the provided cached_dynamodb backend (similar to Django’s cached_db backend).
This probably isn’t a good option for you if…
- You aren’t hosting on EC2. While it’s still possible to use, YMMV as far as response times go. Within AWS, the response times are excellent.
- You’re not at a big enough scale to find yourself with a ton of django_sessions table bloat, or the previously mentioned read/write thrashing in django_sessions.
- You store massive (>1MB) amounts of data in sessions.
If you’d like to give django-dynamodb-sessions a try, either grab it with your choice of easy_install, pip, or other equivalent, or snag it from its page on PyPi. Install instructions can also be found on the PyPi page.
If you think this is interesting…
Make sure to Watch the project on GitHub. If there’s enough interest, I’ll continue to tweak and improve the package.