Experimental Testbed: Each node of our testbed has two 4-core 2.53 GHz Intel Xeon E5630 (Westmere) processors and 24 GB main memory. The nodes support 16x PCI Express Gen2 interfaces and are equipped with Mellanox ConnectX QDR HCAs with PCI Express Gen2 interfaces. The operating system used was RedHat Enterprise Linux Server release 6.1 (Santiago).

These experiments are performed in 8 DataNodes with a total of 32 maps. Each DataNode has a single 1TB HDD. The NameNode runs in a different node of the Hadoop cluster and the benchmark is run in the NameNode.

The RDMA-based design improves the write throughput by 19% - 25% and write latency by 18% - 22% compared to IPoIB (32Gbps). The RDMA-based design also increases the write throughput by 22% - 39% and reduces the write latency by 19% - 24% compared to 10GigE.

Experimental Testbed: Each node of our testbed has two 4-core 2.53 GHz Intel Xeon E5630 (Westmere) processors and 24 GB main memory. The nodes support 16x PCI Express Gen2 interfaces and are equipped with Mellanox ConnectX QDR HCAs with PCI Express Gen2 interfaces. The operating system used was RedHat Enterprise Linux Server release 6.1 (Santiago).

These experiments are performed in 4 DataNodes with a total of 16 maps. Each DataNode has a single 300GB OCZ VeloDrive PCIe SSD. e NameNode runs in a different node of the Hadoop cluster and the benchmark runs in the NameNode.

The RDMA-based design improves the write throughput by 42% - 47% and write latency by 13% - 27% compared to IPoIB (32Gbps). The RDMA-based design also increases the write throughput by 56% - 64% and reduces the write latency by 17% - 31% compared to 10GigE.