ABSTRACT
ABSTRACT As a new trend for data-intensive computing, real-time stream computing is gaining signicant attention in the Big Data era. In theory, stream computing is an eective way to support Big Data by providing extremely low-latency processing tools and massively parallel processing architectures in real-time data analysis. However, in most existing stream computing environments, how to eciently deal with Big Data stream computing and how to build ecient Big Data stream computing systems are posing great challenges to Big Data computing research. First, the data stream graphs and the system architecture for Big Data stream computing, and some related key technologies, such as system structure, data transmission, application interfaces, and high availability, are systemically researched. en, we give a classication of the latest research and depict the development status of some popular Big Data stream computing systems, including Twitter Storm, Yahoo! S4, Microso TimeStream, and Microso Naiad. Finally, the potential challenges and future directions of Big Data stream computing are discussed.
