infoTECH Feature

January 06, 2016

Automating Storage Decisions to Enable Machine Learning with Smart Objectives

By Special Guest
Lance Smith, Primary Data CEO

As one year prepares to turn into another, it’s a good time to take a step back and look at how the deck is stacking up for the future of technology. Times are really changing: we’ve got self-driving cars on the streets of Silicon Valley, and this year, new iPhones were purchased via telepresence.

In the IT industry, technologies like data virtualization are starting to shake up the silos of storage that have caused costly sprawl and resource waste. By leveraging a global dataspace that spans storage types, data virtualization makes it possible to manage data through objectives. As these objectives evolve and become intelligent, they may in fact lead the way to decision based machine learning that helps IT stay a step ahead of global events that take shape in an instant.

Smart Objectives automate decision making so that data can be migrated to the storage type that best meets Basic Objectives for performance and protection.

Get Smart, Data: Objectives for Applications and Across Infrastructure

Basic Objectives assign a specific level of performance or protection needed by an application, whether that’s for example, Platinum, Gold, Silver, Bronze, or any other category defined by the IT team. It’s important to note the application-centric focus here – these foundational objectives manage what enterprise applications need to be effective business tools.

Smart Objectives evolve from today’s widely understood application Service Level Agreements (SLAs). Smart Objectives enable enterprises to maximize efficiency across their entire infrastructure. When application objectives can be better served by placing data on available server flash, shared SAN and NAS, or cloud resources, Smart Objectives make the decision to automatically migrate that data to the available storage type that meets or exceeds the basic objective.

For example, if data is cold for too long, this can trigger the data to be moved or demoted to lower performance storage to free up space for a more active file. In addition, Smart Objectives provide intelligence on application usage that enable informed decisions to be made about which applications to promote to higher objectives. With this reach across different storage types, Smart Objectives focus on the infrastructure as a whole. When these actions happen automatically, expensive data migrations become history, and IT can focus on other strategic tasks that add more value to the business.

With Smart Objectives, companies can be ready for systems to respond to critical events in an instant. For example, when a music celebrity passes away, it often triggers immediate demand for their legacy albums. With Smart Objectives monitoring the infrastructure, it quickly becomes known that the music file has become hot, and then the content can be moved automatically to a fast resource to ensure smooth performance before it triggers a fire drill. Similarly, as a video goes viral around the globe, the content can be moved quickly to meet demand. For more traditional enterprises, this example could be when a promotion is launched and is luckily even more successful than anticipated. The rapid response made possible by Smart Objectives ensures happy customers take their carts - whether real or digital - through the checkouts with less stress for IT teams.

From Insight to Learning

The data-awareness that make Smart Objectives possible can be supercharged by analytics. This makes it possible for systems to “learn” datacenter needs over time.  Once the system can proactively anticipate when demand may exceed storage supply, it can begin to reprioritize resources based on IT policies to ensure critical applications are never interrupted. In addition, it becomes possible to make recommendations on what types of storage are truly needed based on usage patterns.

For example, over time, systems can “learn” that some applications are hot at the end of each month, or each quarter, or even the fiscal year when an entire enterprise completes HR evaluations. As the system begins to understand when these demands happen, Smart Objectives can evaluate which resources are needed and when they are critical, helping companies reduce overprovisioning expenses over time.

Learning Alone is Not Intelligence

It’s important to distinguish machine learning from Artificial Intelligence. Smart Objectives can lead to machine learning, which can also lead to AI, but there is a level of independent “thinking” in AI that is different from machines learning and making recommendations based on past events.  Advanced machine learning may resemble AI, but we’re still a long way from realizing the direct link bridging learning and intelligence.

In the near term of 2016, Smart Objectives will revolutionize storage as we know it by allowing enterprises to focus on what data needs instead of designing dedicated, overprovisioned storage systems for mission- and business-critical applications. With data virtualization delivering a global dataspace, companies will have more choice than ever before. Enterprises will be able to aggregate storage into a global pool to scale both performance and capacity on demand, rather than years in advance of actual needs.

Smart Objectives ensure the right data is placed on the right storage at the right time to deliver efficiency breakthroughs that will save companies millions. Integrating diverse technologies from different vendors will become easy, which benefits both end customers as well as companies undergoing acquisitions or mergers, since these often introduce a struggle to integrate very different datacenters. Being able to achieve this without replacing a single system means that enterprises can maximize existing investments even while adding innovative new technologies to their datacenters.

While the future promises more breakthroughs to come, the innovations on our doorsteps today certainly remind us that the future is now. True Artificial Intelligence may be a concern for many of our era’s greatest thinkers, including Primary Data’s own Chief Scientist Steve Wozniak (News - Alert), who wisely cautions against its use in weaponry. While storage is still a very long way from anything close to singularity, we stand to gain invaluably from machines that can learn and react to meet the Smart Objectives we set for the systems that keep us, businesses, and even our world connected as we enter 2016.

Lance Smith is CEO of Primary Data, who you can follow on Twitter (News - Alert) for more information on data virtualization: @Primary_Data




Edited by Kyle Piscioniere
FOLLOW US

Subscribe to InfoTECH Spotlight eNews

InfoTECH Spotlight eNews delivers the latest news impacting technology in the IT industry each week. Sign up to receive FREE breaking news today!
FREE eNewsletter

infoTECH Whitepapers