Preparing for the unknown: how predictive capacity analytics is changing the face of mobile network planning

17 JANUARY 2017

Radio Access Network (RAN) equipment has historically been the one of largest areas of CAPEX for mobile carriers.  This includes the radios and ancillary equipment needed to provide the actual connections to all our wireless devices.  If you are a network operator, this is critical stuff, and as demand for data continues to grow, it doesn’t come cheap.  For instance, according to AGL Magazine, Verizon Wireless’ 2016 CAPEX spend is estimated at $11 billion, (yes, that’s Billion, with a B), approximately half of which went towards investments in RAN.

This isn’t Your Father’s Network:  Managing New Complexities in the era of 4G
At this level of spend, operators want to be sure they are maximizing resources, delaying additional investments and optimizing current equipment and assets wherever possible. With yesterday’s networks this wasn’t as challenging, but 4G/LTE has introduced new layers of complexity that need to be closely managed, especially as these networks begin to steadily fill their capacities. With 4G there are many ways to manage network capacity – everything from standard optimization practices and deploying small cells, to re-farming network spectrum, partnering with another carrier, or simply moving hardware resources around. To make the right choices, operators need the right tools – and the right information. And with today’s networks, it’s not enough to look to past experiences to guide these decisions.  We need to consider the new options brought by 4G and be able to predict the future as well. Fortunately, we now have the tools to make this happen.

The Capacity Management Conundrum – Why More isn’t always Better
Analyst firm Analysys Mason estimates the typical mobile network is overprovisioned by an average of 25%, just to provide enough leeway to manage unexpected demand and ensure network performance KPIs are met.  This, and the limitations of network planning tools, resulted in a rather simplistic approach to managing network capacity.  While basing investments on simple traffic forecasts may seem like a good idea, it’s also very costly and can still leave operators with pockets of congestion and disappointed customers.  We believe there is a better way.

ROI vs. Traditional approach
The traditional approach to managing network capacity uses yesterday’s traffic and creates a simple, often rough growth forecast, which fails to account the many dynamic aspects of today’s networks. This can lead to bad CAPEX decisions that don’t address the most critical – and potentially most costly – issues.  In the past, there wasn’t much of an option, as network planners only had a limited amount of historical network information and Excel spreadsheets in their toolbox; fortunately, this has improved in recent years.

The Changing Face of Network Planning:  Predictive Network Capacity Analytics
The network capacity planning tools that are needed today should be predictive, analytics-driven solutions that look nothing like tools of the past.  They should utilize rich data sets and advanced forecasting algorithms.  Automated what-if analyses should be performed, and recommendations optimized based on return on investment. This would be a more intelligent, cost-effective approach to managing network growth.

As shown in the image below, network and traffic data should be collected and combined with customer business rules. Then an impact analysis would be performed, and critical, ‘next-best’ actions identified.

This would create a prioritized list of upgrades, but in a timeline-based approach that doesn’t just focus on short term results. It would solve today’s problems but also look to the future, providing mid-term (3 – 12 months) and long-term (1 – 5 years) growth strategies. For instance, short term goals can focus on immediate needs and typically include issues like identifying and resolving network bottlenecks.  For this, there would be analyses of areas that can have a more immediate impact, looking at data related to load balancing, along with equipment features, parameter optimization and perhaps even the need for extra licenses.

For mid-term goals, carriers typically focus on optimizing network growth while also meeting budgeting objectives. In other words, getting the most bang for their buck, so they may want to look at adding (or swapping) hardware, investing in new cell sites, adding new carrier partners or rehoming initiatives.

Long term goals typically plan for growth one to five years out, and should focus on long-term, strategic initiatives.  For these scenarios, carriers might want to explore network sharing, investments in new technologies and architectures, or even spectrum re-farming and the possibility of deploying small cells. (See chart below.) With the right tools in place, carriers can plan for all these scenarios and more, as they move from putting out today’s fires, to creating a comprehensive strategy for ROI-driven growth.

Achieving Results – Key Proof Points
By taking an analytics, ROI-driven approach, TEOCO has helped some of the world’s largest carriers optimize their network with an eye towards the future.

Customer 1:
A wireless operator wanted to enhance its 3G capacity across its 8,000 cell network.  In this situation, we created a strategy for addressing both short and mid-term goals, initially focusing on optimizing the carrier’s network by modifying antennae tilt parameters. This reduced their instances of network overload by 64%. Then we addressed their longer-term budgeting objectives, working to determine how to reduce CAPEX.  This effort resulted in a 30% reduction in the number of new cell sites required, and a 10% savings in their network upgrade budget.

Customer 2:
A wireless operator wanted to improve 2G/3G capacity across a major metropolitan region, comprised of 1,370 cell sites.  In this instance, there were no short-term requirements, but we created a 3 – 12-month strategy that led to 30% fewer base transceiver station upgrades and 39% fewer NodeB upgrades. TEOCO’s solution was more accurate at predicting network traffic, which the service provider had been overestimating by as much as 35%.

ASSET Capacity: TEOCO’S RAN Capacity Management Solution
Due to ever increasing network complexity, radio network design is no longer a standalone business activity. A Capacity design tool needs to connect to the wider OSS/BSS eco-system to leverage the vast amounts of valuable data available.  ASSET Capacity is TEOCO’s network capacity planning tool, providing advanced traffic modeling capabilities allow you to plan current and predict future capacity requirements, meaning network upgrade plans can be implemented before service degradation affects customers, and only where required. “What-if” scenarios help to plan where to best deploy hardware and how to best use different technologies to meet the data explosion challenge.