Your business’s data is growing—exponentially. But your competitiveness depends on your ability to gather, store and get value from that “big” data.

Sources like mobile devices, ubiquitous sensors and social media can tell you things—but only if you can understand what they’re saying.

Here’s how… Big Data

Successful companies find signals in this sea of data: They make informed decisions, which create massive competitive advantage.

But others see no signals. They see only the cost and burden of managing explosive data growth. They’re wasting a business-critical asset.

Buying infrastructure that supports and extracts value from this kind of growth is always challenging.

When leasing a warehouse or office premises, or even buying an IT system that will drive business value, there’s always been an element of risk: How big do I buy? If you buy something too big, you waste money that could have been spent elsewhere. If you buy too small, you can stop business growth in its tracks.

You have to wonder, is there a size that’s just right?

A solution to this is to buy your space in bite-sized chunks as you need it. With office premises, you can take an additional lease on another property and continue the growth there. But this approach can cause problems of its own: It creates management challenges in supporting distributed teams, managing lease arrangements, coordinating security access, and determining which resources and people belong in which office.

IT infrastructure has startlingly similar challenges around security, management and workload placement.

So how should you tackle it? There are two main philosophies:

1. “Scale-Up”

One solution to your office expansion problem could be that you buy or lease a few large buildings to house your employees.

Similarly, when it comes to building IT infrastructure to manage big data, you can use a small number of large computers. These can house the majority of your applications including their sudden spikes in demand.

This reduces management costs, eases security compliance, and makes capacity planning simpler. Each new cycle of purchases allows more applications to run faster, with added hardware power: faster processors, more memory, bigger disks.

This approach is called “Scale-Up.” It’s been the mainstay of computer technology purchasing since the days of the mainframe. blue pane of glass transparent skyscrapers

1. “Scale-Out”

After Hurricane Sandy, stackable disaster relief housing was made from shipping containers. These are often cheap to buy, can be deployed quickly, but also come with a number of drawbacks.

This approach is also seen in enterprise IT, where standardized infrastructure blocks are hooked together. It’s called “Scale-Out.”

It was pioneered to solve big scientific problems like climate modeling. It ties together a whole bunch of different machines into something that could be made to act like one big machine.

They could do this because the problems they were trying to solve were easy to split into thousands of tiny tasks. Gaining the quaint name of “embarrassingly parallel,” these problems could run simultaneously on many different servers, with the thousands of small results gathered and reduced into a single answer. bigdata-real-estate–istock-Spectral-design

1. The Best Of Both Worlds

But there’s a middle path, which combines the best of both the scale-up and scale-out approaches.

Let’s use the building analogy again: It’s similar to building out a business campus. Multiple high-quality buildings, constructed over time, each created to your specifications—some standardized, others more specialized. Site access, security and other services are optimally placed to improve productivity, value and manageability. v

This gives you the best of both worlds: You can add new capacity as required, in increments that make sense when you need it. You can meet initial requirements with very modest budgets, yet easily absorb hypergrowth.

Let’s stretch the analogy some more: You can even throw in some temporary accommodation—“demountables” in the IT world—if you need to, or grab some serviced offices—like spinning up a new cloud instance. Then you can consolidate them back again when the rush is over—all without disruption to your existing applications or productivity.

The Bottom Line

Your business will benefit from analyzing your big data. But first, you have to store it.

A pragmatic approach combines the benefits of scaling up when it makes sense and scaling out when needed. Doing this allows you to forget arcane technical arguments about storage architectures.

That allows you to turn massive data growth into compelling business advantage.