Wide-Area Spark Streaming: Automated Routing and Batch Sizing

Abstract

Modern stream processing frameworks, such as Spark Streaming, are designed to support a wide variety of stream processing applications, such as real-time data analytics in social networks. As the volume of data to be processed increases rapidly, there is a pressing need for processing them across multiple geo-distributed datacenters. However, these frameworks are not designed to take limited and varying inter-datacenter bandwidth into account, leading to longer query latencies.In this paper, we focus on reducing latencies for spark streaming queries in wide-area networks, by automatically selecting data flow routes and determining micro-batch sizes across geo-distributed datacenters. Specifically, we formulate a nonconvex optimization problem, and solve it with an efficient heuristic algorithm based on readily measurable operating traces. We conducted experiments on Amazon EC2 with emulated bandwidth constraints. Our experimental results have demonstrated the effectiveness of our proposed algorithm, as compared to the existing Spark Streaming.