As I understand things, the Scrum backlog is composed of a series of Stories that represent something for the end user and this is further decomposed into Features.

If this is the case, where does all the behind the scenes features go that aren't really linked to a story but are still useful?

For example, say I'm making an application that catalogs the contents of a hard drive. A story wouldn't require it but having an md5 hash on each file would be a nice feature.