High-end IBM storage products lead the industry in both speed and cost, according to recently released test results from the vendor-neutral standards body the Storage Performance Council (SPC).
The benchmarks showed that the IBM Systems
The SPC report, released earlier this week, said the DS8300 Turbo demonstrated the best performance for a single enterprise disk system, outperforming the previous product version, the DS8300 by 22%.
In the SPC-1 IOPS benchmark, the top five performers, ranking from highest to lowest, were:
However, when measuring SPC-1 price/performance, no IBM product made it into the top ten. The price/performance ratings (SPC-1 $/IOPS) were, in order:
"IBM has raised the ceiling in terms of performance, usability and features," according to Walter Baker, administrator for the storage performance council. Baker also said VARs should be looking at the attributes of the new model and how the new features could be applied to create new customer opportunities.
Despite that high level of performance, price and requirements of customers will be critical factors that will weigh on end customer's decisions to buy these products, he said.
"Some customers may not need the level of performance demonstrated by the IBM "high-end" results, given a customer's actual requirements. We encourage end-users to not only look at the "big numbers" generated by large configurations, but to determine what their actual performance requirements are, and then consider results and configurations appropriate to those requirements," Baker added.
IBM took the opportunity to reaffirm its commitment to its approach to storage technology. "Our customers can be assured of excellent performance whether they want to scale up with high-end disk systems like the DS8300 or scale out with a virtualized solution with SAN Volume Controller," said David Vaughn, IBM's worldwide product marketing manager for enterprise disk storage.
The company also challenged other vendors with competitive products to submit their offerings for similar testing.
"Not all storage vendors participate in the SPC benchmark, which makes it difficult for customers to determine whether products from those vendors can meet the needs for their demanding applications," Vaughn added.
According to Baker, the SPC-1 benchmark measurement consists of random I/O operations that have queries and update functions. These include email, reservation systems, sales and data entry systems.
The SPC-2 is a benchmark that measures sequential operations like video on-demand streaming applications and computer-assisted design solutions, such as downloading files from a central server, working on the file and then uploading the file to the central server. It also measures long queries to extract data from a database.
Both benchmarks were used to test and measure the performance of the IBM System Storage SAN Volume Controller and the results concluded that the product gives the best performance for a virtualization solution.
Indeed, in both the SPC-1 and SPC-2 categories the SAN volume controller has produced the highest scored of any technology submitted for testing.
In the SPC-2 benchmark, IBM's products once again topped the class in features and speed as measured by SPC-2 MBPS, but again costs hurt their ranking:
However, when comparing SPC-2 price performance, dollars per SPC-2 MBPS (SPC-2 $/MBPS), the top five products were
IBM System Storage SAN Volume Controller 4.1 ($400.78) was in eighth place.
The SPC is a non-profit corporation with an independent testing body that produces performance results that are both audited and peer reviewed for industry compliance.
Storage vendors that are members of the SPC include Sun Microsystems, Hitachi Data Systems, HP, IBM, NetApp, NEC, 3PAR and Dell. The most noticeably absent vendor to decline having their products tested by the SPC is EMC.