Getting all-flash storage investment approved

We are at an interesting time in the all-flash storage array market. While naysayers still abound, we didn’t have much trouble finding advocates during a recent online survey on the topic. The evidence suggests that the more recently you have looked, the more likely you are to be a fan. This is an indicator of how all-flash options have progressed from an expensive luxury with a limited feature set, to a more broadly viable alternative to traditional HDD systems, over a relatively short space of time.

But while an increasing number of technical staff are enthusiastic, particularly about the step-change in performance and related benefits for IT, the notion of go-faster storage doesn’t, in itself, excite the average business exec in quite the same way. These comments from some of our survey participants sum up the difference in priorities and perspectives:

“Speed is generally a secondary concern, so you need to figure out another way to present the case to management.”

“Performance is great but not a primary motivation [for business execs] in most cases.”

“For our engineering users, it’s all about performance. All-flash is where it’s at. For our business users, it’s all about cost.”

Picking up on this money theme, many believe that there is a good argument to be made around cost savings, especially when you broaden the scope of the discussion:

“Reduce power, space, cooling. Defer datacentre expansion. Better performance delivered for space/dollar deployed. Less time spent in diagnosis of performance problems.”

“Lower cost of management, and lower cost of running, power and cooling.”

“Focus on TCO versus older technology. The real wins are less power, cooling and floor-space, as well as fewer people to manage more storage because of greater simplicity – both in management of the new gear and in the wizarding required to eke out better performance compared to a disk-based system. Also show how you can get more from the other parts of the infrastructure, e.g. less IO wait time on processors means higher VM density/socket.”

That last point about knock-on optimisation can obviously be translated to savings on server equipment, and, often more importantly from a cost reduction perspective, a lowering of software licensing costs.

The cost-case can be strengthened even further if you don’t just increase the scope of your assessment, but also look ahead to consider spend and predictability over the entire lifetime of the investment:

“All-flash arrays having a service life of closer to 7 years makes them feel as if it’s a better ROI. Also, the fixed maintenance deals from some vendors mean it’s easier to budget, especially given the price hikes traditional vendors usually force on you in years 4 and 5.”

“The ability to stretch out an investment over a longer time-frame than just a few years. Moving disks into newer arrays in the future helps to appease the bean counters, as it’s not dead money.”

“Predictable, linear cost for growth and maintenance over the solution life, with increasing discounts as overall spend increases on the solution.”

“Flash has a longer duty-cycle than disk, so there are savings over the longer term.”

What’s coming through quite strongly here is a consistent message that if we look beyond direct, immediate acquisition costs, it is possible to build a strong cost-efficiency case for an all-flash array investment. That’s not to say, however, that you need to rely on this alone. Other participants in the survey suggested that if you articulate benefits in business terms, this can have an impact too. The potential resilience-related advantages were highlighted by some, for example:

“Non-IT execs tend to be far more concerned about uptime than security or performance.”

“I am starting to see more focus on resiliency. With a lot of the big name major outages, I can justify more spend if resiliency is somewhat guaranteed.”

“Explaining the value to the business (or consequence of not obtaining it). Things such as downtime due to existing arrays. If you can attach a $ value, then that will help.”

There’s then the sustainability angle if that’s something your execs care about:

“Reduced requirements for space and power mean that flash is also a green solution.”

But arguably the most powerful type of investment case, if you are able to make it, relates to driving the organisation’s top line as well the bottom line. This is where you start to impact the numbers to which many exec bonuses are linked – always useful for focussing minds:

“In one recent example, it was easy because we could reduce the time to market for an application release, which accelerated time to invoice by 50% (through reduced development cycles, less infrastructure, and faster performance). The return on the $400K investment in flash delivered $3.2m in additional revenues in the fiscal year, and lowered DC costs for the platform by 15%.”

“Show execs what´s in it for them in terms of business impact. Let´s say all-flash means faster time to market, less downtime / better uptime, that means more business transactions per period without interruptions. As another example, a Web application click-to-response time without flash could be 3 seconds, and with flash could be about 1 second. With a 100,000 clicks per day, the impact is huge.”

“Make it about business outcomes. We can reduce a rack of storage to 6U and consume a fraction of the power, and that’s great in IT terms, but what that actually means to the business is a reduction in datacentre costs by X% a year. Improving performance by Y% then translates to improving our customer satisfaction rating from 7 to 8 over a 12-month period.”

All good arguments, but we all know that however solid the logic and strong the case, some execs don’t have the time or patience to even try to understand:

“No evidence based approach particularly elicits any interest.”

One response is to adopt an approach that has helped achieve a couple of recent surprise referendum and election results – keep hammering people with a few simple messages, even if you need to put the messy facts to one side, and dumb-down complex issues to one-line sound-bites relating to the things that are particularly close to their hearts:

“Nearly all execs have a short attention span and can’t be bothered with the detail. As such, ’convincing’ them is usually down to some combination of repeating ’this is cheaper’, ’this will lower costs’, etc, over and over again. At the end of the day, facts are largely immaterial.”

“Don’t baffle them with numbers, just a simple message saying without buying ’it’ we can’t continue trading, we won’t meet our contractual commitments, etc – that’s what works.”

Of course we could never officially condone such behaviour, but sometimes it’s a case of doing what’s necessary to get what you want. If you are dealing with an exec who just can’t cope with the detail, here are a couple of other respondent suggestions on particularly simple messages to try to get the point across:

“Imagine how crap your iPhone would be if it had a hard drive in it. Now times that crappiness by the number of users you have. That’s effectively what we are living with at the moment.”

“Faster helps everyone. By today’s standards, hard disks are now like floppy disks.”

“Flash is already the present, not the future. If you’re not on board already, you’re a future archaeological find of an extinct species.”

Maybe that last comment is a bit extreme, but the truth is that if non-tech execs are standing in your way on all-flash solutions, they are probably stopping you taking advantage of other technology advances too, so likely extinction may well not be that far off.

Zooming out to the bigger picture, while many participating in our research are clearly flash enthusiasts, the reality is that all-flash arrays will sit alongside HDD and hybrid systems for some time to come in most environments. Technology choices always need to be made in light of past investments, existing skills sets, incumbent supplier arrangements, and so on. Furthermore, there are some applications where the flash cost-benefit equation still doesn’t quite cut it. Long term archiving, is an obvious example here.

If you are looking to invest in all-flash, however, some of the ideas laid out above will hopefully have got you thinking on ways to further your agenda internally. You might also find our full research summary interesting too, particularly when exploring options and negotiating terms with suppliers. You can download your copy of this by clicking here.