The current Internet technology has already entered the big data, artificial intelligence phase, a variety of computing models, calculation engine after another. From the MapReduce distributed computing began 10 years ago, almost every three years there will be a new calculation engine change, from the pedigree of Hadoop, stream computing and later Storm, after the emergence of memory-based iterative calculation of Spark red pole moment, and then again to Flink potential lightning speed today into our field of vision. At the same time just data computing, machine learning framework in recent years is endless: Keras, PyTorch, Caffe2, TensorFlow ...... can say that every rise of a calculation engine, mean computing technology a new breakthrough.
With the development of computational engine, and now the Internet business services increasingly rely on big data and artificial intelligence technology, from the initial analysis of the data warehouse-based business, it has now evolved into a more real-time, higher complexity of intelligent decision making.
A combination of both to bring more business innovation possible, but the technology behind the phenomenon is increasingly large and complex technical systems,