In-Depth
4 Data Center Predictions for 2013
What are the Next Big Things data center managers are watching?
By Gary Orenstein
As 2013 looms on the horizon, companies planning for the year ahead have IT teams exploring the Next Big Thing impacting data centers. Here's what they're looking at.
Prediction #1: Software Will Define the Data Center
We have witnessed the coming-of-age for software-defined networking, marked most prominently by VMware's acquisition of Nicira and continued theme towards software defined data centers. Companies now use software to perform functions that previously required specialized hardware. This shift -- the evolution from specialized hardware to industry-standard platforms and software -- has happened before, and it will happen many more times. Smartphones represent one of the most important examples. Today's smartphones combine calculators, cameras, GPS units, and (of course) a telephone into one device, all by using software.
In 2013, we'll see the rise of software-defined storage. Using software, companies can now turn nearly any server into performance shared storage. Combining powerful flash memory and software allows the server to appear as a shared storage device. Although shared server storage probably will not immediately replace legacy disk arrays for capacity, it provides the acceleration needed to ease the load on overburdened data centers, reducing the load on existing storage resources and extending the useful life that investment.
Expect businesses to continue looking for ways to make the data center perform more efficiently through innovative software, enabling them to do more with less.
Prediction #2: Big Becomes Bigger
Big data is now less of a mystery and more of a staple in the everyday data center. Companies realize how big data insights can help them better understand their customers' needs and how to reach them.
The enormity of big data has challenged companies to find solutions that can meet their big data demands within realistic budgets. However, there is more to big data than size alone. As IBM notes, big data spans four dimensions: volume, velocity, variety, and veracity. New application acceleration solutions such as flash memory and software-defined storage help meet both the capacity and speed requirements for big data analytics, ensuring real-time and relevant results delivered within budget.
Recently, big data has been used to catalog endangered species, help developing countries, and let people find their data doppelgangers, but this is just the tip of the iceberg. "Data scientist" is now a bona fide corporate position, underscoring the scope of this new business opportunity. Solving complicated big data problems will happen much more frequently in 2013. Even as today's big data becomes tomorrow's norm, other datasets will continue to grow. Leaders will continue to seek faster and more powerful systems to keep up once they see the true potential of big data solutions, all because innovators will continue to ask bigger and more complex questions.
Prediction #3: Applications Will Boldly Go Where No One Has Gone Before
I predict that in 2013 software developers will begin to take advantage of new primitives and protocols developed for flash memory. This will allow developers to treat flash as a new memory tier that is more akin to DRAM than storage. As a result, enterprises will be able to program applications to utilize terabytes of NAND flash as memory, greatly accelerating application processing speeds by extending the capabilities of system memory.
By using flash specific APIs, such as those created by my employer, Fusion-io, applications will no longer need to rely on primitives and protocols from the disk drive era. This native access to flash memory will provide dramatic results for companies looking to gain an edge over the competition by getting answers faster than ever before.
Prediction #4: Virtual Barriers Will Break Down
Many companies are already virtualizing their applications, but in 2013, more companies will turn to virtualization than ever before. Gartner estimates that 60 percent of server workloads will be virtualized by 2014, which seems to be in line with what we've seen at Fusion-io.
As servers become more powerful, virtualization will become more prevalent as enterprises look for ways to take advantage of these compute resources. These powerful new systems will also pave the way for companies to virtualize even the most I/O-intensive applications. Using powerful caching tools, companies will be able to realize the full potential of virtual environments.
Breaking With the Status Quo
When you look at these predictions, you will see that software-defined data centers link them together. Leaders in 2013 will combine innovative technology to solve the big data demands of our modern world. In addition to capturing more market share, expect to see increasing efficiency (such as smaller and more nimble data centers) and better customer experiences. This balancing act cannot happen with conventional IT infrastructure, and 2013 will make that more apparent than ever.
Gary Orenstein is the senior vice president of products at Fusion-io. Prior to Fusion-io, Gary served as vice president of marketing at MaxiScale, a developer of Web-scale file systems acquired by Overland Storage. From 2006 to 2008, Mr. Orenstein was the vice president of marketing and business development at Gear6, offering memory-based caching and file acceleration solutions. Mr. Orenstein holds a B.A. from Dartmouth College and an MBA from the Wharton School at the University of Pennsylvania. You can contact the author here.