Analytics and the Next-Generation Network

By Michael Pawlak, Senior Industry Consultant Communications, SAS

When we talk about network generations, it’s not the same as device generations. Device generations are discrete—2G becomes 3G becomes 4G. Network generations are fluid, taking a more hybrid approach to accommodate legacy technologies. Consider the transition from rotary to touch-tone dialing, or the fact that 3G users still outnumber 4G users, who often have to rely on 3G infrastructure.

So the next-generation network—we’ll call it NGN for brevity’s sake—is not just about 5G mobile technology, though that will be a cornerstone for reasons we’ll discuss below. NGN will be about anticipating the demands of emerging technologies and applications, and ensuring the network can deliver them a high-quality experience.

NGN will also be about incorporating emerging and established technologies to improve network performance. Cloud computing allows scalability and versatility; a network customer might have a device on her desk, but the actual functionality exists in the cloud, centralizing changes. Software-defined networking virtualizes routing and switching functions, allowing reconfiguration on the fly. And advanced analytics ties it all together, making sense of huge pools of data and automating the application of business and network process optimization.


5G mobile technology, arriving between 2018 and2020, can handle 50 times the traffic of 4G technology. And that’s critical because of the volume of data created and collected by the network. Consider that a single phone call creates hundreds of data points from network elements like routers and switches—and that voice traffic is no longer the primary use of the mobile network; data traffic is.

That’s just baseline. The number of mobile subscriptions will almost double to almost 7 billion in the next three to four years. But an even more significant challenge to the network than user interactive data traffic comes in the form of machine-generated traffic (M2M)—the Internet of Things (IoT).

The range of Internet connected devices that are invisible to the mobile user is staggering, and the possibility of connecting more is there all the time. Today’s car may have as many as 54 electronic sensors for engine operation, brakes, safety features, alignment and more. Connected to the Internet, these sensors could alert a driver to a problem, provide directions to an appropriate auto shop—augmented by an IP-connected GPS—and book the next available appointment. Machines in factories now routinely have electronic sensors to monitor their health and performance so preventative maintenance can be scheduled. In several U.S. cities, sensors monitor the presence of cars to manage parking. Cisco Systems and General Electric estimate there will be a trillion sensors in the world by 2020.

In 2014, there were 5 billion Internet-connected M2M devices, according to Machina Research. That number will be 27 billion by 2024. In fact, it could be more as new applications are developed to take advantage of the trillion-sensor world.

Advanced analytics is necessary to make sense of and apply this lake of data to issues of network configuration, traffic management, customer service levels, and more.


At Rogers, the Network Quality Group has had to run reports on the data from the equipment of hundreds of technology vendors, trapped in their proprietary shells—manually, a process that could take weeks. Advanced analytics can turn that task into a near-real-time experience.

Analytics is about answering questions; advanced analytics can ask them autonomously. Some of the typical NGN issues that advanced analytics can address:

  • Traffic management. In the IoT world, some traffic can go east-west without going north-south to a data centre. Some sensor technology can communicate with peers without having to pass through the middleman. Analytics can help determine what the most effective routing is and apply it on the fly.
  • Dynamic network configuration. In conjunction with software defined networking (SDN) and network function virtualization (NfV)—which put the backplane on commodity hardware—advanced analytics can monitor types of network traffic by region, and reconfigure to, for example, rebalance voice traffic against data traffic.  Advanced analytics can also help with the scheduling of planned outages—and also ring a warning bell if they’re causing unplanned outages.

Advanced analytics can also contribute to the planning of hardware and software purchasing by flagging geographic and application demand.

  • Customer service level management. By applying analytics to customer network usage, the telco can automate the process of optimizing rate plans for customers and creating value-ads on top of the service. Analytics can also sort high-value and low-value customers, making sure network resources are distributed appropriately.

Once, a telco’s primary offering was a voice network. In the world of the NGN, the primary offering is the network. Voice, data, video—all these will be essentially value-adds on top of it. If it’s done right, that will be invisible to the customer. And that is what advanced analytics is for: Getting it right in real time.

Get More Insights

Want more Insights from SAS? Subscribe to our Insights newsletter. Or check back often to get more insights on the topics you care about, including analytics, big data, data management, marketing, and risk & fraud.