This nginx module provides fixed boundary range requests to request large files as well as remapping random range requests to fixed ones providing effective caching in nginx.
Currently this module requires the range_filter.patch be applied to the nginx
core. Hopefully this will be upstreamed to nginx so this step won't be needed
in the future.
Using nginx as a reverse caching proxy is great but when the first client requests a file it must be downloaded in its entirety from the origin before nginx can start to send it on to the client. For small files this isn't an issue but for large ones the time taken to receive the file from origin significant, large enough for the client to time-out waiting for the initial response.
This module uses range requests to avoid this issue, downloading the file in small chunks which can start to be sent to the client as soon as each chunk is received.
Some clients that use range requests do so on random parts of the file meaning that caching these parts provides little or no benefit.
This module uses fixed boundary range requests to avoid this problem.
This module provides a new nginx config option range_cache_size which configures
the size and boundary of range requests sent to the origin, overriding any existing
range request and splitting non-ranged requests.
For each sub request that this module uses to request from the origin it configures
the nginx variable $range_cache_range which can be used as the cache key component
to avoid request conflicts.
A good example of how to use this module can be found in Multiplay LANcache.