Top-n recommendation technology has recently received a lot of attention in information service community.In this paper

we study the problem of top-n recommendation under the cloud data.Firstly

we propose a multilayer distributed storage architecture(MDSA)which can fit the need of top-n recommendation.Secondly

based on the consideration of reducing the cost of network transmission

we present the data encoding pattern which is based on the MDSA architecture.Lastly

we fast realize top-n recommendation using the map/reduce distributed programming model.For satisfying the realistic requirement

we further present three application variants of top-n recommendation.Massive experiments demonstrate that our approach is efficient and has high overall performance.