Analytics in IT Operations - Defining the roadmap
(Published on behalf of Arun Kumar Barua)
As the speed of business accelerates, it is a lot more critical to have visibility into IT operations. However, getting that information in a form that you can use to direct faster and more informed usable decisions is a major challenge. Visibility into operations is one thing but turning massive amounts of low level, complex data into understandable and information useful intelligence is another. It must be cleansed, summarized, reconciled and contextualized in order to influence informed decisions.
Now let's think about it, what if organizations are able to effortlessly integrate their data; both structured and unstructured data within the organization? What if it were easy and simple for businesses to access it all? Think of a situation where this data acquisition process is predictable and consistent! Business insights is linked to a framework for quick decision-making and made available to all who require it.
In a previous post, we looked at the importance of the data that is generated on a daily basis through IT operations. Recognizing the importance of these data and analytics is essential but putting in place the processes and tools needed to deliver relevant data and analytics to business decision-makers is a different matter.
Predictive analytics is not about absolutes, it encompasses a variety of techniques from statistics and the use of machine learning algorithms on data sets alike to predict outcomes. Rather, it's about likelihoods. For example, there is a 76% chance that the primary server may failover to secondary in XY days. Or there is a 63% chance that Mr. Smith will buy at a probable price, or there is an 89% chance that certain hardware will be replaced in XY days. Good stuff, but it's difficult to understand and even complex to implement.
It's worth it, though. Organizations that use predictive analytics can reduce risk, challenge competitors, and save tons of money on the way.
Predictive Analytics can be used in multiple ways, for example:
- Capacity Planning: Helping the organization determine hardware requirements proactively and a forecast on energy consumption
- Root Cause Analysis: Detecting abnormal patterns in events thus aiding the search exercise on single-point of failures and also mitigating them for future occurrences
- Monitoring: Enhanced monitoring for vital components which can sense system failures and prevent outages
The selection of an apt tool will enable you to use reports and dashboards to monitor activities as they happen in real-time, and then detail them into events and determine the root-cause analysis to realize why it happened. This post talks a bit more about the selection of such a tool.
By identifying various patterns and correlations with events that are being monitored, you can predict future activity. With this information, you can proactively send alerts based on thresholds and investigate through what-if analysis to compare various scenarios.
The shortest road to meaningful operational intelligence comes by generating relevant business insights from the explosion of operational data. The idea is to transform from reactive to proactive methods to analyze structured & unstructured operational data in an integrated manner. Without additional insights. it is likely that IT management will continue to struggle into a downward spiral.
Now would be a good time to tap into the data analytics fever and turn it inward.
(Arun Barua is a Senior Consultant at Infosys with more than 9 years of diverse experience in IT Infrastructure, Service and IT Process Optimization. His focus areas include IT Service Management solution development & execution, Strategic Program Management, Enterprise Governance and ITIL Service Delivery.)