Network Managers:
Time to Escape From Alarm-Based Diagnostics

Life isn’t easy these days for telecom companies, who are facing the fast-growing need for new investment, in 5G and IoT technologies, in core fiber and in updating older equipment in general.

As the Boston Consulting Group said recently, “For telecom companies, undertaking a digital transformation is especially challenging. These businesses are usually large and complex. They have legacy products and IT systems. And they have cultures and skill sets that often harken back to a pre-digital era. Yet, telecom companies have an especially strong need to transform.”

Life isn’t easy for network managers, either, who will be challenged to support these changes while working hard to maintain customer loyalty in the face of ever-more-heated competition. They’re dealing with customers made more fickle through greater opportunities to switch network vendors, and who become even more disgruntled as initial promotional prices give way to higher bills.

Keeping Customers Happy

Should the network manager, whose traditional responsibility involves keeping the infrastructure running, be worrying about holding and attracting customers? Absolutely. And that’s a major change from the practices of the past.

Another major change is that network managers have a promising new ally in the fight for improvement: analytics, or more specifically IoT analytics.

But what exactly to do with these analytics, and how to do it, is the tricky part. There will be petabytes of new data coming from the proliferating IoT probes and devices; the better the telecoms and their managers leverage this data, the more likely they’ll enjoy robust growth in the years ahead.

How can IoT analytics help? Look for solutions to problems such as these:

  • Problem – Network alarms are too localized to convey higher-level information about issues or outages. False positives can overwhelm operations and thus delay reaction times to the real issues. Add to this the fact that networks take in more data and grow more complex and cumbersome by the day, so the problems worsen.

o Solution – IoT analytics can analyze granular event populations in real time to detect nuanced problems and reduce false positives via advanced anomaly detection.

  • Problem – Problem – Network complexity slows the efforts by humans to analyze problems that extend across multiple data sets.

o Solution – Visualization tools, machine-learning algorithms and predictive analytics help managers perform root-cause analysis with speed and accuracy.

  • Problem – Legacy networks often rely on temporary workarounds in order to minimize the incident impact, and manual repairs to resolve the issues.

o Solution – IoT analytics quickly initiate repairs and/or workarounds and continue to monitor impact in real time to assure the network, and the users, of full restoration.

  • Problem – Customers expect fast recovery and immediate, even proactive, communication about service outages.

o Solution – IoT analytics add a customer-service dimension above the network plane. This dimension combines extensive information on network operations and customer knowledge with the network’s alarm profiles. This real-time view through customers’ eyes can help speed remediation, while giving the network the ability to communicate instantly with affected customers.

Starting At The Top

IoT analytics can help transform network operations in many ways. Ultimate success with analytics, however, requires major cross-network cultural changes, and these changes must start with a company-wide commitment at the top of the organization.

One sign of such a commitment involves structural changes that give higher-level status to customer service and to the IT resources that support these services.

Another sign is a commitment to an end-to-end analytics infrastructure that can service the company’s business units while preparing for the larger transformations to software defined networking and digital operations.

Such an infrastructure should, actually must, be cloud-based.

Today’s cloud solutions can easily handle the performance needs of billions-per-day IoT events as well as the calculations of the various machine-learning and other models that add value to the data flow. Cloud computing is elastic, so it can harness whatever computing resources are necessary for the job, but then shrink when computing demands lessen. On-premise equipment is the opposite: you’ve got to install more compute power than you need for headroom. On-premise equipment adds greatly to capital costs, and, of course, guarantees obsolescence.

As for the analytics infrastructure, an end-to-end architecture that can be used by different groups pays dividends in efficiency and agility over the alternative, which is to use a variety of “point” analytics products.

Standardizing Data

For one thing, an end-to-end analytics architecture maximizes efficiency and security by helping you standardize data definitions, security controls, and even look-and-feel of reports and dashboard screens. This benefits inter-group communication and C-level appreciation for the insights being generated by different groups.

For another, it lets you take advantage of reusable software via building-block applications, allowing certain analytic functions to be used by multiple groups as part of a common component library.

At a higher level, an end-to-end architecture puts everyone on the same page, so to speak. This is valuable today and will become even more important as telecoms make their way into the beckoning worlds of full digital operations.

Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong>