I'm looking at eXtreme Programming and its release iterations based on User Stories. But I wonder how to integrate huge internal processes into this particular mind-set.
For example, the Google crawler doesn't relate to a specific User Story, it index web pages in order to make them available to the User Story "As an Internet User I want to provide key words In order to find web sites". But the development of a Crawler is a huge set of work that isn't atomic and can't be done in a single iteration.
How can this kind of development be handled by this methodology?