Scaling any algorithms to very big input is a part of every onsite Google interview. Could you please share any materials/links about it? The only approach I know currently is map reduce but it's not suitable for all problems.