dcsimg

Hitachi Reopens Benchmark Debate

Download the authoritative guide: Enterprise Data Storage 2018: Optimizing Your Storage Infrastructure

SHARE
Share it on Twitter  
Share it on Facebook  
Share it on Google+
Share it on Linked in  
Email  

Hitachi on Monday claimed record-setting performance for its high-end Universal Storage Platform V, reopening a long-simmering debate over the value of storage industry benchmarks.

The results — 200,245.73 SPC IOPS in the Storage Performance Council's SPC-1 benchmark using just a single storage controller — were the highest ever for an enterprise storage system in the SPC-1 program, nearly doubling IBM's December benchmark for its high-end DS8300 Turbo.

Chuck Standerfer, senior partner at the Evaluator Group, called the results "very impressive," and encouraged other enterprise storage vendors — namely, EMC — "to graciously accept this challenge and publish their SPC results so that customers would have a valid independent source for comparisons."

EMC, for its part, reiterated that it doesn't participate in SPC benchmarks, "as they are not indicative of real-world environments or an objective assessment of a system's capabilities handling the random workloads that are common in customer environments."

For example, EMC said vendors "can turn off features and functions to maximize their performance during the tests that would never be turned off in a real-world application. When quoting SPC results, vendors do not have to indicate what they disabled. So vendors can turn off all RAID function, turn off all HA functions, turn off all data integrity functions, etc., in order to look good and that does not provide useful or beneficial information to anyone except the vendor."

EMC's views were also the subject of a blog by company executive Chuck Hollis in July.

Standerfer agreed that "there is no such thing as a perfect benchmark test, including the ones that vendors use internally. However, I believe that this one is representative of reality, and provides a good independent means of comparisons. If the SPC is not 100 percent representative, the variance would be the same for everyone."

The USP V also leads the pack in maximum cached IOPS with 3.5 million peak cached IOPS, 75-300 percent better than other tested systems.

For more on the SPC-1 results, visit the SPC Web site.

Back To Enterprise Storage Forum

Submit a Comment

Loading Comments...

NewsletterSTORAGE DAILY

Want the latest storage insights?