The development of Institutional Repository requires a massive resource accumulation, the demand for storage keeps increasing especially for the small files. This article proposes a distributing storage model Hdspace which is based on Dapace and HDFS to resolve the problem of the storage of massive small files of Institutional Repository. First by a means of merging small document files to get new storage files, then by establishing two indexes based on subjects and index pre-caching to improve the file-reading response, finally put forward a method for the storage of massive small files.