Leveraging ‘Big Data’ analytics to boost IT department’s operational resiliency and efficiency

BANGALORE, INDIA: Businesses use structured data analytics to understand business trends, predict sales, guide investments, and control risks. IT uses analytics to manage Web stores, plan production capacity, and highlight data security concerns.

During the past five years, structured data analytics -- data mining, data profiling, data forensics, risk analysis, and predictive modeling -- gained prominence in IT. Analytics now takes its place alongside applications, transaction processing, business intelligence, and content management.

In any business, the expectations from IT are a lot more. Business processes and the IT systems that support them need to go inevitably faster is an expectation at any organization.

The reality is that businesses and IT run in cycles and that operational speed needs to be able to change to match those cycles. It is more important for processes and systems to go the right speed, which is not always the fastest speed. Building operational resiliency is a business goal that affects the requirements for and use of IT infrastructure and applications.

With the use of Analytics, IT can deliver more efficient services, improve resolution time, and better align its operations with the IT infrastructure and IT service management objectives.

The key is in identifying opportunities to improve the effectiveness, efficiency and productivity of current event/incident management operations; and validate the customer benefits of proactive event management. A more aspirational end-to-end event management approach will in this case be necessary in an IT services context.

Major improvements like the following can be seen in the IT Operations with the use of an analytics approach:

1. Reduction in the size of the incident volume that helps IT to:

• Prioritize assets more likely to fail: Reduces event volume to monitor by about 70 percent
• Focus on event spikes: Reduces signals that need to be monitored by additional 20 percent

2. Prediction of incidents better and earlier and lets IT to intervene proactively to:

• Use Big-Data based "industrial process control" approach: predicts up to 70 percent to 95 percent of incidents using event signals, up to 15 to 40 minutes earlier

3. Increases IT team's time spent on proactive event management and:

• Reallocates capacity away from non-event /incident work: 10 per cent to 20 per cent additional time for proactive event management is created with this

• Improves incident management efficiency and reallocates capacity to event management: ~10 percent additional time for proactive event management is also created.

4. Develops end-to-end approach for event management across the stack

• Focuses on service uptime: Gives the opportunity to consolidate application and infrastructure monitoring, improve process integration across the stack etc

Enterprise business is complex. Big data Operational Analytics promises to manage this complexity to make better decisions. But the technology services that run your business are also complex. Many are too complex to manage easily, fueling more complexity, delays, and downtime.

To combat this onslaught, you can no longer just accelerate current practices or rely on human intelligence. You need machines to analyze conditions and to invoke appropriate actions. These actions themselves can be automated.

IT Operational analytics is an exciting field because it represents breakthrough innovations that can bring substantial value and lasting competitive advantage to the businesses that adopt them and the vendors that create the right solutions.

Service management process improvements are a prerequisite to any other effort, including the tools. To perform adaptive, full-service automation, you need IT analytics, a disruption to your existing monitoring and management strategy.

(Murale Narayanan is Director, Global IT, EMC Corporation and Tarun Sareen heads India IT CoE & APJC Theater Lead, EMC IT, EMC Corporation)

(The views expressed in this article are that of the author and do not necessarily reflect the views or policies of CIOL)