Implementing an Async Redis Cache Backend for Django

Based on the available information, implementing an async Redis cache backend for Django would be a valuable addition to the framework's asynchronous capabilities.

Current Context

Django introduced asynchronous views and middleware support in version 3.1, allowing developers to create non-blocking request handlers. However, as noted in the release notes, Django's ORM, cache layer, and other components that perform long-running network calls did not initially support async access.

In Django 4.0, a Redis cache backend (django.core.cache.backends.redis.RedisCache) was introduced, providing built-in support for caching with Redis. This implementation is currently synchronous.

Implementation Approach

Your suggestion to add a new class that leverages the native async support provided by the redis package aligns well with Django's approach to async features. Here's why this approach makes sense:

1. Consistency with Django's async philosophy: Django has maintained backward compatibility by introducing new async-specific components rather than modifying existing synchronous ones.

2. Leveraging native async capabilities: The redis-py package already provides async support, so creating a dedicated async class would allow direct use of these features.

3. Performance benefits: True async implementation would avoid blocking the event loop during Redis operations, which is particularly valuable in async views that make multiple external calls.

Implementation Considerations

If you proceed with implementing an async Redis backend:

1. Class structure: Consider creating an AsyncRedisCache class that parallels the existing RedisCache but uses async methods.

2. Compatibility: Ensure it works seamlessly with Django's async context management, which automatically handles transitions between sync and async code.

3. Configuration: The implementation should use the same configuration pattern as the existing Redis backend, making it easy for users to switch between sync and async versions.

4. Testing: Include comprehensive tests for both ASGI and WSGI environments to ensure proper functionality.

This approach would add some code duplication, but as you noted, the Redis backend is relatively small, so maintenance overhead should be manageable. The benefits of true async support would likely outweigh these costs.