What math and the price of flash have in common
One of the reasons I ended up an editor is that I'm pathetically bad at math; I still count on my fingers, and it takes me an embarrassingly long time to do simple problems in my head. But one thing that's been affirmed in my time covering technology is that math is really hard.
Take determining the true cost of flash, for example. Objectively, solid-state drives are more expensive than hard disk drives, so flash is more expensive than spinning disk, right? Nope. There are a lot of variables companies need to consider when looking at flash pricing: dollars per IOPS versus dollars per gigabyte, added data reduction techniques, the needs of the workload, the life of the drive and more. That makes determining whether flash is worth the spend a lot harder than figuring out what time a train leaving Chicago will arrive in Detroit.
Another complicating factor of flash pricing is that vendors -- of HDDs, hybrid arrays and all-flash systems -- all say that their product is the best value. And if there's another thing I've learned from covering tech, it's that vendors are great at bending numbers to reflect what they want. They're not necessarily lying, but the mileage an organization gets out of an array will probably be different than what a vendor promises.
My trick to avoid doing math -- which is to act like I'm thinking about the problem, but then just wait for someone else to come up with the solution -- doesn't work for tech buyers trying to get a handle of flash pricing, so shop smart. Consider all of those dollar figures to determine the real cost of flash drives. And don't forget that issues with NAND supply affect the flash market. If you're the one tasked with justifying the price of flash, there are real variables to consider.
