“Parallel builds specialized crawlers to access content from the "deep web," such as data behind portals that traditional search engines do not index.”