How CloudFront processes partial requests for an object (range GETs) - Amazon CloudFront
Use range requests to cache large objects

How CloudFront processes partial requests for an object (range GETs)

For a large object, the viewer (web browser or other client) can make multiple GET requests and use the Range request header to download the object in smaller parts. These requests for ranges of bytes, sometimes known as Range GET requests, improve the efficiency of partial downloads and the recovery from partially failed transfers.

When CloudFront receives a Range GET request, it checks the cache in the edge location that received the request. If the cache in that edge location already contains the entire object or the requested part of the object, CloudFront immediately serves the requested range from the cache.

If the cache doesn’t contain the requested range, CloudFront forwards the request to the origin. (To optimize performance, CloudFront may request a larger range than the client requested in the Range GET.) What happens next depends on whether the origin supports Range GET requests:

  • If the origin supports Range GET requests – It returns the requested range. CloudFront serves the requested range and also caches it for future requests. (Amazon S3 supports Range GET requests, as do many HTTP servers.)

  • If the origin doesn’t support Range GET requests – It returns the entire object. CloudFront serves the current request by sending the entire object while also caching it for future requests. After CloudFront caches the entire object in an edge cache, it responds to new Range GET requests by serving the requested range.

In either case, CloudFront begins to serve the requested range or object to the end user as soon as the first byte arrives from the origin.

Note

If the viewer makes a Range GET request and the origin returns Transfer-Encoding: chunked, CloudFront returns the entire object to the viewer instead of the requested range.

CloudFront generally follows the RFC specification for the Range header. However, if your Range headers don’t adhere to the following requirements, CloudFront returns HTTP status code 200 with the full object instead of status code 206 with the specified ranges:

  • The ranges must be listed in ascending order. For example, 100-200,300-400 is valid, 300-400,100-200 is not valid.

  • The ranges must not overlap. For example, 100-200,150-250 is not valid.

  • All of the ranges specifications must be valid. For example, you can’t specify a negative value as part of a range.

For more information about the Range request header, see Range Requests in RFC 7233, or Range in the MDN Web Docs.

Use range requests to cache large objects

When caching is enabled, CloudFront doesn’t retrieve or cache an object that is larger than 50 GB. When an origin indicates that the object is larger than this size (in the Content-Length response header), CloudFront closes the connection to the origin and returns an error to the viewer. (With caching disabled, CloudFront can retrieve an object that is larger than this size from the origin and pass it along to the viewer. However, CloudFront doesn’t cache the object.)

However, with range requests, you can use CloudFront to cache an object that is larger than the maximum cacheable file size.

Example

  1. Consider an origin with a 100 GB object. With caching enabled, CloudFront doesn’t retrieve or cache an object this large. However, the viewer can send multiple range requests to retrieve this object in parts, with each part smaller than 50 GB.

  2. The viewer can request the object in 20 GB parts by sending a request with the header Range: bytes=0-21474836480 to retrieve the first part, another request with the header Range: bytes=21474836481-42949672960 to retrieve the next part, and so on.

  3. When the viewer has received all of the parts, it can combine them to construct the original 100 GB object.

  4. In this case, CloudFront caches each of the 20 GB parts of the object and can respond to subsequent requests for the same part from the cache.