How To Get Real-Time Analytics In Synch With 5G

prodcastfour_blog7 With the networking industry’s move to 5G, today’s telecommunications providers face a set of challenges more intricate and complex than ever before.

On the one hand, telecoms are, or will soon be, able to offer transformative network-related performance that goes far beyond anything that’s been possible in the past. On the other, telecoms are themselves having to transform in order to respond to the new customer demands – and competition – that will come from 5G.

By now, with all businesses facing the challenges of harnessing “big data,” and with telecoms looking at the additional challenges of operations transformation, it’s clear that some form of real-time analytics capability is essential to future success. But most telecoms (and businesses) don’t have a solid vision of just how to get the most out of the many analytics tools now available.

Telecoms should start with a general template that will give them the flexibility and power to synchronize their analytics efforts across the company and out to their supplier and customer ecosystems. This template will support the two most important functions of analytics organization and performance: breaking down operations silos and maximizing agility of analytics applications.

Breaking Down Silos With An End-To-End Infrastructure

There may be a temptation to try a variety of different analytics tools in different parts of the organization, but that would work against the thing that transforming-to-digital businesses need most: internal operations harmony.

Today’s telecoms are trying to do many things at one time. They’re merging their OSS and BSS systems and preparing to migrate to a core 5G architecture; they’re upgrading and adding new virtualization capabilities, such as NFV (network function virtualization) and network slicing, to prepare for more lucrative services-oriented contracts; and they’re trying to keep their customers satisfied, even as new features – and complexities – are added to machines and devices by their makers.

The good news is that, with virtualization, the necessary operations functions to support customers will be in software rather than hardware. But peel back the surface and you‘ll see that lots of software interaction is required to achieve a fairly simple result.

For instance, a video-on-demand service may require 10 or more underlying applications to interact seamlessly, from the smart-device application itself down to the software needed for compression and decompressions, for handling adaptive bit rate, for routing and switching and for compute and storage needs. Each of these areas may represent a type of silo, depending on the organization’s structure and the state of its 5G transformation.

And end-to-end analytics infrastructure can bridge these and other silos by extending a layer of open connectivity across the organization and even reaching out to customers and suppliers.

In the video-on-demand example, the real-time analytics infrastructure brings in streaming data from the different processes and applications involved, analyzes them to create a normalized score, and then watches for anomalies. If a customer device goes down, the analytics can dive down quickly to find the root problem.

To make the link from customer down to network, the analytics may have to traverse any number of operations silos. But the result is telling. Without such analytics, network operations, IT and customer-support might waste their time pointing at one another while the customer is waiting for service.

To do its work, a real-time analytics infrastructure will be able to access data from across the organization, searching databases and data lakes for information that can be used as context for new analysis. This, too, helps break down silos – namely, the departmental databases that may have been inaccessible to outsiders in the past. Thanks to the analytics infrastructure, this information can now be used to add context – on everything from past customer behaviors to machinery repair histories – that can inform new analysis.

Maximizing Agility With Low Code Development Tools

No one knows what new 5G-inspired applications may turn up in the future. Yes, we see the obvious ones, like robots and smart cities and virtual reality and self-driving vehicles. But those will just be scratching the surfaces of innovations that will come looking for “big” connectivity.

Savvy telecoms will build their analytics teams, which will typically include data scientists and network and business analysts, and then will want to give these teams as much flexibility as possible to support new business development with analytics applications. Low-code, developer-friendly tools can help maximize teams’ agility in a number of ways.

For one, dashboard-based building-block designs help technical and non-technical team members to communicate easily without getting lost in technical fundamentals. For another, they promote reusability, so code blocks developed for one application can be quickly customized to be used with others, thus speeding time-to-build. And these tools will be compatible with popular machine learning modeling languages such as Spark and R, thus helping speed development of analytical models.

Finally, the better development tools will include fast-start templates that are specialized for specific operations, such as anomaly detection or predictive analysis. Templates can hold basic algorithms, data models and UIs for, say, anomaly detection, which might save 70 or 80-percent of the effort to get started. Then the user adds solution-specific functions such as the unique data structures and workflows for their application.

The results are new analytics applications that come up in days or weeks, rather than months or years. And that capability, along with the silo-breaking benefits of an end-to-end infrastructure, will help telecoms make the most of their analytics resources as they move up to the brave new world of 5G networking and applications.

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>