Last week, industry arbiter The Storage Performance Council (SPC) announced a new benchmark for measuring the energy consumption of storage devices.

Called the SPC-1E benchmark, the test is an extension of SPC Benchmark 1C (SPC-1C) for storage components. It uses the SPC-1C workload and incorporates the complete set of SPC-1C performance measurements and reporting. Additionally, it adds the dimension of energy consumption at differing operating points and is available for end-users and original storage manufacturers that can now use the benchmark to predict energy efficiencies.

As solid state drives (SSD) become more prevalent, energy use becomes an even more critical issue in price-performance comparisons. In addition to the core SPC-1C performance measurement and reporting requirements, the SPC-1C/E benchmark extension defines two states of benchmark execution: idle and active (performance).

The benchmark implements multiple idle phases to demonstrate power management/savings features, configuration and instrumentation requirements for energy use measurement and execution requirements for idle state measurements. In addition, the benchmark measures the transition into active (performance) state measurements, data collection requirements for energy use measurements and expanded disclosure and audit requirements to address the new energy use measurements and reported data.

IBM was one of the first vendors to use the SPC-1E benchmark to test the performance and power consumption of its IBM System Storage EXP12S, an array that uses SSDs. In a configuration of a IBM POWER6 server, 1.5GB cache SAS RAID adapters and the EXP12S, benchmark results showed 45,000.20 IOPS, with an energy consumption of 121.31 IOPS per watt. The IBM EXP12S results demonstrate the value of the new SPC-1C/E benchmark and its ability to measure both throughput (IOPS) and energy consumption.

Deni Connor is principal analyst for Storage Strategies NOW and host of both the Masters of Storage and Masters of Servers Solution Centers.