The present invention relates to a Hadoop ecosystem, which comprises: a distributed message queuing device for classifying stream data for each topic and temporarily storing the same; and an application server for fetching data for each topic from the distributed message queuing device at a predetermined time in accordance with a scheduler policy, and distributing the fetched data to each platform. According to the present invention, by distributing the stream data to each platform in the Hadoop ecosystem based on a scheduler, it is possible to reduce duplicate codes and improve productivity of the entire system.
展开▼
