By Prem Chand, CEO and President, Milestone Technologies, Inc
According to IBM’s 2014 analytics study based on its survey of more than 1,000 businesses worldwide, the value driver for big data has shifted from volume to velocity. Now that great breakthroughs have been made in managing seemingly infinite types of big data, the big data leaders are the businesses that offer their clients the ability to interpret and act on meaningful information faster than competitors. The study* also states that executives in Asia and the Americas dramatically increased demand for action-oriented data-driven insights throughout 2014, and the trend is expected to sharply increase and spread across the globe in 2015.
Speaking from the perspective of an IT services provider, this trend calls for a re-imagining of Network Operations Center (NOC) management as a platform for delivering engineering services. Increasing the speed with which skilled engineers can take action based on insightful data is the single most important key performance indicator (KPI) of any NOC. It all boils down to one simple idea: automate as much of the process of converting raw data into actionable data as possible. This kind of automation greatly reduces human error and frees up engineering resources. This re-architecting of a traditional NOC allows highly skilled employees to focus solely on insight-driven problem solving with speed, precision, and agility.
NOCs produce large volumes of raw data that come from multiple sources. This ocean of data contains drops of deep insights into every aspect of a given network. However, no one would filter all of the water in the Pacific Ocean by hand. Automated data analysis helps NOC management in much the same way that massive desalination machines aid the management of emergency drinking water systems. Such intelligent automation allows NOC professionals to focus on preventing problems and handling emergencies, instead of spending their precious time on tedious mind-numbing tasks.
Data analytics processes must be intelligently aligned to business processes. That is the basis of a context-driven network. The idea of a NOC as a platform for engineering services is rooted in the idea that network operations are most effective when they are optimized to enable engineering goals. In this case, the primary goal is identifying threats to a network and providing actionable data to engineers capable of neutralizing those threats. Smart automation meets that goal the instant data related to such threats can be detected. Intelligent automated systems are most effective when they have access to strategically implemented problem detection software. The following guidelines outline best practices for effective problem detection.
Best Practices for Problem Detection:
Link automated data analysis systems directly to problem detection software, so that data can be analyzed the instant a problem (or potential problem) is detected. Use proactive problem detection software that recognizes patterns and trends associated with previous problems, instead of systems that search only for known problems. Automated and detailed problem logging is critical to creating proactive problem detection systems because they document those patterns and trends that led to past problems. Ultimately, the most important consideration may be through categorization. The only way to isolate actionable data as rapidly as possible is to prioritize the most serious problems detected on your network. In this way, you can solve problems that could create outages without delay. Again, automating the detailed documentation of problem logs can be of great help in recognizing new categories of problems and raising awareness about which problems can lead to outages.
Best Practices for Data Analysis:
The best automated data analysis systems act like data detectives, investigating problems and isolating root causes in record time. Employing these systems eliminates time delays and human errors that occur when engineers and technicians are charged with meticulously combing through massive amounts of data. Instead of spending countless hours manually searching for actionable data, problem managers can start off with detailed information about a problem, its root causes, and information about how similar problems have affected the network in the past. Smart automation practices can even make it possible for managers to see all of this information in a report that:
- prioritizes current problems,
- isolates actionable data, and
- provides a clear picture of urgent next steps.
In this paradigm, NOC engineers and technicians can dedicate all of their expertise, time, and energy to taking strategic actions necessary to prevent service failures, instead of frantically trying to repair networks and restore services.
The value of velocity in big data management is undeniable. The smartest thing to do is to make the path to actionable data as short as possible in your organization. This is especially true when it comes to mission-critical operations, such as preventing network failure. My recommendation to every executive charged with an enterprise that generates big data is to take stock of every bottleneck in your organization and work with knowledgeable people who can optimize networks for velocity.