Original Articles
Kong Xiangsheng and Chen Jianb
Abstract
Scientific workflows produce large amounts of scientific data. Hadoop MapReduce has been widely adopted for data-intensive processing of large datasets. The Kepler system can support scientific workflows, high–performance and high-throughput applications, which can be data-intensive and compute-intensive. The paper presented a "Kepler + Hadoop" framework for executing MapReduce-based scientific workflows on Hadoop