Advert Label?
In future El Reg, could we have a label on an article when it's just an advertorial rather than a genuine article?
When you have an existing storage array infrastructure with a variety of server apps about to hit the array, how do you know if array technology upgrades or even a new array will work as well or better than the existing kit? Do you trust your vendor and generalised performance data: It does 450,000 IOPS? This version is 1.3 …
So how do you propose that the Reg inform its readers of new and useful products? I found the article interesting and plan on sharing with my team.
But maybe I missed something... Can you point out where and how the author crossed the line?
I learned a long time ago to evaluate the info conveyed; is it interesting? is it valuable? is it something I might use? is it good for the community? Did the author trash anyone? Does the author have a history of being unfair? If anything the author did right by masking the identities of the tested; one product won over the other for a specific workload, but who knows how the winner would fare in a different workload.
I've watched customers run scripts that drive up IO to the max and make purchasing decisions based on the max performance; and have been frustrated when you try to explain that max IO is not an indication of the real world.
This tool is a potential next gen way to understand how a product would work in a customer's specific environments; definately a step above a swingbench.
I see no foul...
And as Sgt. Friday used to say, "just the facts mam"; the author did just that.
Take a look at the Linux utility blktrace, as a $0 alternative to this.
It is written by the same author as the "fio" I/O test software. One of the nifty features is that blktrace can save the I/O pattern using a format that fio can re-play. This is ideal for capturing real life workloads to test against different storage systems.