There are two HBase micro-benchmarks: (1) Put Latency and (2) Get Latency. For detailed configuration and setup, please refer to our userguide.
Put and Get Latency
Put Latency
Get Latency
Experimental Testbed: Each compute node in this cluster has two twelve-core Intel Xeon E5-2680v3 processors, 128GB DDR4 DRAM, and 320GB of local SSD with CentOS operating system. Each node has 64GB of RAM disk capacity. The network topology in this cluster is 56Gbps FDR InfiniBand with rack-level full bisection bandwidth and 4:1 oversubscription cross-rack bandwidth.
These experiments are run with 1 Regionserver and 1 Client. For each message size, the client creates a new table and performs 10000 operations on the table. For the Get latency test, we put one row in the table and get the same row for each operation. The RDMA-IB design improves the latency of Put and Get operations by up to 88.4% and 81.1% over IPoIB (56Gbps), respectively.
