Abstract: For the storage and retrieval problem of library works in the era of big data, this paper implemented the massive storage of library works with HDFS and used ElasticSearch to distributed index and retrieval. Test results show that under the large amount of data, this system needs about onetwentieth retrieval time of the traditional Oracle database for searching the same search words, and it has the cache function. So an efficient and distributed digital library retrieval system is built.

Key words: ElasticSearch, digital library, mass storage, distributed index, distributed retrieval