Case Study: Transforming Legacy Telecoms Networks

It is no exaggeration that telecommunications operators worldwide retain an abundance of ‘legacy’ networks: those using decades-old technologies for which support and maintenance contracts, software updates and hardware parts have already ceased to be available.

Legacy networks become increasingly expensive to maintain as they age: a dwindling source of parts requires pricey refurbishment of the old, a situation exacerbated by accelerating failure rates causing network and service outages, and even liquidated damages to be paid.

These networks should have been retired long ago. However, that they still garner significant revenue, directly and indirectly from the millions of services and other networks they transport – business voice, data, mobile access and core, emergency services, control and signalling – such that continuing worth demands some sort of technology transformation.

After all, proprietary and dated tools, and manual processes associated with them, can be transformed alongside, to technologies such as Software-Defined Networking (SDN) and virtualised networks that are highly automated.

So, what stands in the way of that transformation? The cost of maintaining the legacy network should outweigh the cost of transforming them, but it is not that straightforward unfortunately. Cost, risk and feasibility prove to be a very complex and circular interaction, and that is what has held back such investment, even by the most resourceful of operators.

3 Problems

Three factors dominate their dilemma:

1. Employees familiar with legacy technologies and their arcane proprietary management tools, are a diminishing proportion of the workforce. As they retire year on year, that undermines confidence, to the extent that the problem is thought best left alone
   
   
2. Service and billing records and the actual network configuration - the so-called back-office - is data generally only in partial agreement with each other, incomplete, and not always an accurate reflection of reality. Sometimes this data is not available – older nodes can fail management communications – or are in difficult-to-consume formats. Without a reconciled and complete view, no one really knows if transformation is feasible, let alone how to conduct it reliably.
   
   
3. Selecting the starting point is critical to success, but even with a clear big-picture strategy, so many detailed considerations and constraints contrive to make this far from obvious. Evaluating many, occasionally opposing, tactics and a myriad of interplays (customer, control, in-building services, physical distributions and virtual protections), must be confected – almost magically – into an effort, spend and recovery efficient roll-out that also mitigates all risks.

The Challenge

A large NA telecom local and long-distance operator had an established business case and strategy for transformation, but no longer had a planning team with the modelling capability to do so. Their scheduled goal was behind by years, so they sought to source an outside ‘Planning Tool and Service’ and select parts of their network to which it should be applied to meet their priorities.

LightRiver, a well-established services supplier with advanced monitoring and management tools already deployed in their network, were awarded the contract. 

“Despite our accurate inventory of circuits and assets, we needed a partner that could process tens of millions of lines of data, and build a system to manipulate, sequence, and display the data in a way that was consumable and actionable. Cambridge MC was the perfect partner for us. Their tools and dashboards allow us to change the project sequence depending on the customer’s specific needs in each different area of the network.” 

– Matt Briley, SVP Global Sales & Solutions, LightRiver

The Approach

Our first step was to dispense with the original piecemeal focus on parts of the network, and analyse the whole: big data for deep insights. That revealed ‘simple’ transformations: those without ramifications for other regions, services or networks, and thereby avoid creating a large backlog of implementation work.

That simplicity had to be quantified, to be credible and satisfy the operator’s priorities. We invented a novel ranked evaluation methodology to combine circa 25 complex and often diametrically opposing metrics. This yielded stepwise transformations that were well (but not critically) sequenced, such that dismantling the network became progressively simpler.

Our Data Science and ML were also used to combine back-office records with actual network configuration data from LightRiver’s netFlex platform, reconciling information and filling in blanks, to provide for the first time an accurate and complete view to direct implementation and mitigate risks.

Our automated ‘planning’ process could be conducted in whatever scope, scale and sequence of priorities the operator needed.

Outcomes

The plans produced enabled the operator to:

1. Discover empty resources that could be powered down without any procurement.
   
   
2. Determine the value of recoverable parts, that turned out 5x greater than anticipated, including previously untrackable inventory.
   
   
3. Determine opportunity clusters like whole-site transformations, avoiding repeat site visits boosting field engineering efficiency.
   
   
4. Recover their schedule to the extent that legacy products earmarked for 2025 could be conducted in 2024.