Segmented Caching

Fastly’s Segmented Caching feature allows you to cache resources of any size. Segmented Caching works by breaking resources into smaller segments in Fastly’s cache then recombining or splitting these resources to respond to arbitrary size byte Range: requests from clients. Once enabled, Segmented Caching improves performance for Range: requests and allows Fastly to efficiently cache resources of any size.

How Segmented Caching works

When an end user makes a Range: request for a resource with Segmented Caching enabled and a cache miss occurs (that is, at least part of the range is not cached), Fastly will make the appropriate Range: requests back to origin. Segmented Caching will then ensure only the specific portions of the resource that have been requested by the end user (along with rounding based on object size) will be cached rather than the entire resource. Partial cache hits will result in having the cached portion served from cache and the missing pieces fetched from origin. (Requests for an entire resource would be treated as a byte Range: request from 0 to end of resource.)

Once Fastly has all of the objects necessary to respond to an end user’s request, the Segmented Caching feature will assemble the response by concatenating or pulling portions of objects. The requests back to origin, also called “inner requests,” will have a true value for segmented_caching.is_inner_req and requests from end users, also called “outer requests,” will have a true value for segmented_caching.is_outer_req. If a request is made for an object without segmented caching enabled, both variables will have a FALSE value.

Limitations and considerations

This feature has the following limitations and considerations you should take into account:

• Segmented Caching is not compatible with object compression and ESI. To use either of these features, you must ensure Segmented Caching is disabled.
• Segmented Caching is not compatible with Fastly's Image Optimizer (IO). If IO is enabled, Segmented Caching is disabled automatically.
• HTTP chunked transfer encoding between Fastly and origin isn't supported. Your origin server must frame responses to Range: requests with the Content-Length header.
• HTTP headers If-Modified-Since and If-None-Match are not supported. Fastly only returns HTTP 200 OK and HTTP 206 Partial Content success status response codes.
• URL purges must be authenticated. Segmented caching allows you to purge all range objects for the resource by URL purge, but authentication for URL purge needs to be enabled due to its underlying implementation. Make sure you’ve provided an authorization token for URL purges as described in our purging documentation.
• Segmented Caching cannot be enabled based on resource size. The VCL code to enable Segmented caching must run before the resource is requested from cache, so it is not possible to know how large it will be. However, it is possible to make segmented caching conditional upon the URL (e.g. /video/) or file extension (e.g. *.m4v), and this is a common use case for resources which will benefit from it. Our instructions below contain an example of how to enable the Segmented Caching feature based on extension.
• Resources cached prior to enabling Segmented Caching are not used. If you are enabling Segmented Caching on an existing service, whole resources already in cache are ignored and Varnish will go back to origin to build range request objects. You can choose to purge these or ignore them to be aged out of cache.
• Your cache hit ratio (CHR) will appear lower than it actually is because only outer requests are used in the calculation. For example, if there is a request for the first 100 MB of a resource but Fastly only has 99 MB of 100 MB in cache, the entire request will be counted as a miss in the CHR stat. In practice, only 1 MB had to be fetched from origin and you experienced 99% origin offload.

Enabling Segmented Caching

Use the following steps to enable Segmented Caching.

1. Determine which resources should use Segmented Caching.

   TIP: We recommend focusing on a set of file extensions or a well-defined URL structure that distinguishes the resources.

2. From the All services page, select the appropriate service. You can use the search box to search by ID, name, or domain.
3. Click the Edit configuration button and then select the option to clone the active version. The Domains page appears.
4. Click the VCL Snippets link. The VCL Snippets page appears.
5. Click the Create your first VCL snippet button. The Create a VCL snippet page appears.
6. In the Name field, enter an appropriate name (e.g., Enable segmented caching).
7. From the Type (placement of the snippets) controls, select within subroutine.
8. From the Select subroutine menu, select recv (vcl_recv).

9. In the VCL field, add a VCL snippet that sets the req.enable_segmented_caching VCL variable to true in vcl_recv. For example, to ensure proper caching of the large resources you've identified that contain MPEG-2-compressed video data, you could add this VCL snippet in vcl_recv:

   1
   2
   3
   4
   # my custom enabled Segmented Caching code
   if (req.url.ext == "ts") {
      set req.enable_segmented_caching = true;
   }
   

   This snippet tells Fastly to look for requests for files with the ts extension and then enable Segmented Caching for those files.

10. Click Create to create the snippet.
11. Click the Activate button to deploy your configuration changes.