Goals and Solutions of Data Allocation in Data Center

Nowadays, data centers have become important resources to store a huge amount of data and to offer data access service. In a data center, data are stored in servers, which also perform computing tasks. Usually, initial data need to be transferred to the data center before it starts the service. This paper investigates how to allocate data from external data sources to servers in a data center, particularly exploring the impact of partitioning data. Three goals for data allocation are presented, i.e., minimizing the number of active servers, minimizing the average number of partitions per data, and balancing servers' workload. Different solutions for these goals are proposed to show the tradeoff. Performance evaluation demonstrates that the goals can be achieved well by the solutions respectively.