Abstract:

With the development of Web and wireless sensor networks, the appearance of big data has brought a great challenge to data query and process. The parallel Skyline computing issue of a moving query object under the MapReduce framework was studied. To solve the problem of load balance, a method based on angle partition was adopted. Then a filtering method was introduced to improve the efficiency. On the basis of those methods, a static Skyline algorithm and an event tracking algorithm named MR-Track were implemented respectively. Finally, many experiments were made to verify the effective and efficient of MR-Track algorithm.

Key words: Skyline query, MapReduce, Hadoop, moving object, continuous query