Applications of aggregation for information summary have great meanings in various fields. In big data era, processing aggregate function in parallel is drawing researchers' attention. The aim of our work is to propose a generic framework that enables to map an aggregation function into an efficient massively parallel algorithm that can be executed on modern large-scale data-processing systems. We describe our preliminary results regarding classes of symmetric and asymmetric aggregation functions that can be mapped, in a systematic way, into efficient MapReduce algorithms.