Jury Still Out, But Fusion-io Making Progress

At the same time that it reported 556% higher revenues for its fourth fiscal 2011 quarter, which ended on June 30, Fusion-io announced it will pay approximately $95 million to acquire IO Turbine, a provider of caching solutions for virtual environments based in San Jose, Calif. At closing, Fusion-io, a vendor of solid-state storage and high-performance I/O solutions, will add approximately 20 IO Turbine employees to its engineering team, with an additional five employees across other areas.

The company says it has partnered closely with IO Turbine for a number of months, and that the acquisition "is a key part of Fusion's strategy to enable enterprise customers to increase the utilization, performance and efficiency of their data center resources and extract greater value from their information assets." Fusion-io goes on to state that integrating ioMemory and IO Turbine improves return on investment by increasing the number of virtual machines per physical server and making it feasible to virtualize data-intensive applications that were previously difficult to migrate to virtualized environments while maintaining performance.

However, while revenues soared year-over-year, they were up just 7% quarter-over-quarter, and the company said next quarter's revenues should be lower than the previous quarter's. Revenues for the year were $197 million, up 445% from fiscal 2010, while net income was $4.6 million, compared with a loss of $32.5 million a year ago. For fiscal 2012, revenues are expected to grow 40% to $276 million.

The acquisition was a big surprise, says analyst Jim Bagley, SSG-NOW, who has been following IO Turbine since before it came out of stealth (the company was founded in December 2009 and unveiled officially in April 2011). He says the announcement was made the same day that STEC, a vendor of solid-state drive (SSD) technologies and products, announced its caching software, which is a little broader-based than IO Turbine and has some nice features when using direct-attached storage. STEC's host-based caching software implements SSDs as a performance cache for all commercially available SSD devices (including SAS, SATA, Fibre Channel and PCIe) to accelerate enterprise applications running under Windows, Linux or VMware operating systems.

"IO Turbine is focused on very tight VMware integration [and] the acquisition should help Fusion-io provide more functionality in a packaged way with their product as more and more servers are using SSDs," says Bagley. This will help the company's software business, but it doesn't address the company's biggest challenge--flash management on the CPU, he adds.

"Fusion-io has arguably the best software on the server in terms of heavyweight drivers," but it has not done the kinds of things on PCI support that its strong competitors have done. "They really need to do a hardware refresh in that area," Bagley says.

The other potential hurdle to growing revenues 40% next year is the potential for the flash memory makers to enter the SSD market. "The foundry owners are looking very hungrily at the kind of gross margins they're getting. I fully expect them to enter the market, which will put pressure on Fusion-io's margins," he says.

At the end of April, Enterprise Strategy Group's Steve Duplessie blogged that, as a maker of SSD PCIe cards, Fusion-io has lots of competition and lives at the mercy of the actual flash makers. "If I were in the market for such a product I could call LSI, OCZ, TMS, Virident and 20 others," wrote analyst Howard Marks. "Should Micron, Intel, Samsung or other flash makers want to play in this space, they'll have a cost advantage over Fusion-io due to vertical integration."

Unlike Duplessie, Marks is not ready to write off Fusion-io as a niche product that addresses I/O latency if, and only if, that problem can be addressed within a single server. Once again, the old horse player's motto applies: "Different horses for different courses."

See more on this topic by subscribing to Network Computing Pro Reports Parallel Memory Optimized (subscription required).