"On demand" caching that kicks in only when requests arrive simultaneously. - View it on GitHub