For IT leaders looking to make the case to move their entire infrastructure to flash storage, Wikibon CTO and cofounder David Floyer has a slide deck for you.

Floyer’s new 30-slide presentation goes deep into the economics of flash, looking far beyond the cost of the media and digging into productivity gains that will sweep across the all-flash organization. Stop waiting to make the move, he says. The time for action is now.

Underlining his case is the recent rapid decline in the cost of NAND flash that has brought it to near parity with disk on a raw capacity basis. For performance-dependent applications, flash is already a clearly superior choice, and Floyer’s 10-year cost forecast sees raw flash storage prices falling below disk prices by 2023.

But that’s not the reason to make the move. The two biggest cases for flash are its vastly superior performance and its shareability. Floyer estimates that a reduction in IT response times from three seconds to .3 seconds doubles user response speed and can double productivity. Such response times are simply impossible using magnetic disk.

But the killer app is sharing. Because disk-based databases can’t be accessed by multiple applications simultaneously, organizations must create extracts and copies for each discrete purpose. This not only drives up storage costs but limits functionality because analytical applications must use outdated copies of production data.

Floyer estimates that 90% of the bits on disk arrays in a typical data center are copies of production data. Large and expensive caches are needed to serve up this information on a timely basis, which adds to cost, floor space requirements and complexity. And that’s not even counting the handicap of using outdated information.

Flash doesn’t have this limitation. It uses space-efficient snapshots of live production data to feed each application that needs it. In essence, an all-flash array can serve one copy of the data to dozens of applications at the same time. And because copying is eliminated, the data is more current and thus more useful.

Estimating the real cost of storage means taking into account packaging, power, cooling, maintenance, space and the data de-duplication and compression benefits enabled by flash storage. Floyer estimates that the four-year cost per terabyte of NAND flash will fall from $470 today to nine dollars and 2020, a decline of 98%. Disk won’t stand still, but its cost decline will be much slower, reaching a four-year cost per terabyte of $74 in 2020. That means that five years from now flash will be 88% less expensive than disk.

But why wait? The Wikibon CTO says that moving to an all flash infrastructure today will deliver:

• A six-fold reduction in storage cost from data sharing and copy elimination;
• A four-fold reduction in data capacity requirements from compression and de-duplication;
• A 24-fold reduction in raw storage requirements;
• Much faster response times for all applications, which directly impacts user productivity; and
• The ability to deploy new applications mixing online transaction processing and inline analytics. In some industries, the benefits of this capability are almost incalculable.

Flash will change the way organizations use information because it enables parallel processing of transactional, analytical, operational and archive systems, Floyer asserts. In effect, this is Big Data in real time. The benefits of being able to detect equipment breakdowns before they occur or upsell customers at the point-of-sale are far more compelling than the already compelling savings in storage costs.

But even on a total cost of ownership basis alone, the time is now to move to an all flash architecture, Floyer asserts. He cites examples of a British financial house that has more than doubled the number of production databases in use as well as the productivity of its developers since moving to an all-flash architecture last year. A US electronics distributor combined all workloads on flash arrays and saw a 30% jump in revenue with no additional headcount in just 18 months. And a US software vendor implemented 100% flash for continuous development and tripled its output of software updates.

What did they all have in common? “They all removed the disk boat anchor,” Floyer says. Download his slide presentation in PDF format to see how you can do the same thing. Read more of David Floyer’s exhaustive analysis on Wikibon Premiumin.