Along with today's data explosion and application diversification, a variety of hardware platforms for data centers are emerging and are attracting interests from both industry and academia. The existing hardware platforms represent a wide range of implementation approaches, and different hardware have different strengths. In this paper, we conduct comprehensive evaluations on three representative data center systems based on BigDataBench, which is a benchmark suite for benchmarking and ranking systems running big data applications. Then we explore the relative performance of the three implementation approaches with different big data applications, and provide strong guidance for the data center system construction. Through our experiments, we has inferred that a data center system based on specific hardware has different performance in the context of different applications and data volumes. When we construct a system, we can take into account not only the performance or energy consumption of the pure hardwares, but also the application-level characteristics. Data scale, application type and complexity should be considered comprehensively when researchers or architects plan to choose fundamental components for their data center system.