A Reliability Enhance Method towards Hadoop Architecture
Jia Miao, Guo You Chen, Kai Du, Zhong Jun Fang · Applied Mechanics and Materials · 2013
1PB or more of the data has been the concern of the reliability of the data in large data warehouses. It strongly depends on many interdependent system parameters, such as the replica placement policy, the number of nodes and Previous work has substantially and respectively discussed the individual effects of these parameters, and rarely provide its optimum value, there is no mention of their optimal combination. In this paper, we propose a new object-based repair Markov model. The analysis of this model based on three popular replica placement strategy, we calculate the personal best value of these parameters in the first, then work out the best combination of GA. Compared with the existing models, our model is easier to solve, and to achieve a more comprehensive and practical conclusions.