Building a fault tolerant framework with deadline guarantee in big data stream computing environments

摘要: Abstract Big data stream computing systems should work continuously to process streams of on-line data. Therefore, fault tolerance is one the key metrics quality service in big computing. In this paper, we propose a tolerant framework with deadline guarantee for called FTDG. First, FTDG identifies critical path graph at given throughput, and quantifies system reliability graph. Second, allocates tasks by aware heuristic scheduling mechanism. Third, online optimizes task reallocating vertices on lower response time reduce fluctuations. Theoretical as well experimental results demonstrate that makes desirable trade-off between high low objectives environments.