Sign up for our newsletter and get the latest big data news and analysis.

Taking a Horizontal Approach to Big Data for Better IT and Business Outcomes

Raja Mukerji, President of ExtraHop.

Raja Mukerji, President of ExtraHop.

In this special guest feature, Raja Mukerji of ExtraHop Networks talks about how IT Operations Analytics (ITOA) has started to emerge as a framework for a more horizontal approach to Big Data in IT. Raja Mukerji is the co-founder and president of ExtraHop Networks, a leader in real-time wire data analytics for IT operational intelligence.

It needs to be said: Big Data isn’t a solution any more. It’s a problem. The root of this problem is the fact that both vendors and enterprises alike have gotten mired in the how of Big Data, and have lost sight of the why. As it became possible to collect, store, and analyze ever-greater quantities of data, somehow more became better – more quantity, more sources, faster speeds. Central to too many Big Data initiatives is the gathering masses of information – information that is then thrown into a repository in the hope that some alchemy will produce actionable insight from vast swaths of indiscriminately collected information. A huge garbage can isn’t “better” than a little one if the real solution is to create less garbage.

Enterprise IT is not immune to this problem. If anything, it’s become increasingly susceptible to it as organizations have become more and more reliant on technology to support every aspect of business operations. If something goes wrong, IT finds itself locked in a war room, racing against the clock to troubleshoot the problem before supply chains are disrupted, customers are lost, or patients suffer. For IT, the promise of Big Data has been to give them the visibility and insight they need to more rapidly remediate these problems. All too often, it’s still falling short of that promise.

The Problem with Silos

One of the central tenets of Big Data is that enterprises should be free to correlate, explore, visualize, and query data from a variety of sources; however all too often, fragmented data sets are silo’d into a given vendor’s user interface or visualization tool. This practice limits data freedom, restricts customer choice, and inhibits the creation of valuable insight that comes from correlating a variety of sources.

This is especially true in IT, where many Big Data technologies have promised a single pane of glass into the IT environment … and ended up delivering a single glass of pain as IT discovers the limitations of that silo’d data set. Take machine data (log files), for example. With log analysis and visualization platforms, machine data can help IT teams more easily identify overburdened machines, plan capacity, and perform forensic analysis of past events; however, because machine data is self-reported, it can also be suspect. Trusting a faulty component for self-diagnosis isn’t always a sound strategy. If you have no other data with which to verify what the machine is telling you, it’s hard to know whether you’re getting the truth.

IT Operations Analytics: The Silo Buster?

Over the past few years, a new discipline, IT Operations Analytics (ITOA), has started to emerge as a framework for a more horizontal approach to Big Data in IT. Rather than a single pane of glass, ITOA advocates for the use of a variety of data sources optimized to deliver specific visibility, and the correlation of those data sets to provide actionable insight for IT.

Four key sources of data have emerged as central to ITOA architectures: machine data, wire data, agent data, and probe data.

  • With log analysis and visualization platforms, machine data can help IT teams more easily identify overburdened machines, plan capacity, and perform forensic analysis of past events.
  • Wire data delivers real-time observed metrics across tiers, including critical operational details, and is invaluable when it comes to improving application performance, availability, and security.
  • Agent data is most valuable to developers and QA teams, allowing them to capture method entries and exits, memory allocation and free events, and stack traces.
  • Probe or “synthetic” data provides service checks that indicate to IT when there is a problem.

Each of these data sets provides a different and valuable lens into the health of the IT environment. When correlated for a horizontal view across the infrastructure, they provide the most comprehensive understanding of the performance, availability, and security of the environment, from infrastructure utilization to end-user experience.

Beyond IT: Transforming IT Into a Strategic Business Driver

While the central mandate of ITOA is to take IT Big Data and make it actionable for IT, the horizontal, silo-busting approach it employs also has the potential to transform IT from a support system for the enterprise to a key driver of business success.

Every Big Data initiative should begin with the core question: “What actionable insights can I derive from this data to improve the business?” For IT, this means transitioning our thinking from how data will reduce mean time to resolution (MTTR) when online payment processing is slow, to how that data can help prevent future outages or even improve customers’ online experiences. It means evolving beyond troubleshooting Citrix latencies in a hospital’s VDI infrastructure, to understanding how the same data flowing over the network can equip clinicians and hospital administration with the insight they need to improve patient care and outcomes.

All too often, individual data sources are viewed as exclusive to particular segments of the business. That data lives within specific tools, and is often leveraged by only one team. Just as great business leaders break down organizational silos in order to drive collaboration and productivity across the business, data leaders should be focused on breaking down data silos in order to deliver actionable insight on multiple fronts. By employing this horizontal approach, not only to data sets within IT, but across the entire organization, we can begin to transform Big Data from a problem into a solution once again.

 

Sign up for the free insideBIGDATA newsletter.

Leave a Comment

*

Resource Links: