Testing for the mobile service lifecycle – how to adapt testing to suit your mobile network with Solver.

Testing standard interfaces in 3G, 4G and 5G networks is essential, but only testing with real subscriber data, based on your live traffic and real environment can deliver the insights you need to really model and validate network performance and behaviour. Testing must be adapted to the demands of your unique network.

When you launch a new feature or service upgrade, you need to validate that it performs correctly. To achieve this, a number of factors count. Is the solution compliant to the relevant standards? Does it impact existing behaviour? How does it perform under different demand scenarios? Is there any impact on adjacent nodes and solutions?

This takes time and effort – and, it’s a continuous process, as operators are shifting to DevOps cycles, with much more frequent updates and releases. Ensuring performance delivery in new virtual environments and moving from test to production requires the right test solutions. But that’s not enough.

In this context, modelling what actually happens in the network really matters, not just testing a feature or interface to a specification. What are traffic volumes? What mixture is expected to be encountered? What services and devices are used by your customers? How frequently are specific services and features invoked? You can’t really test performance without reflecting the real traffic mix in your network – which is different from operator to operator and may also differ from one location to another.

As a result, test solutions need to be adapted to the specific circumstances and unique traffic flows in each network. And, to really count, they must reflect actual subscriber data, so that performance can be compared with real traffic profiles.

Solver allows you to accomplish this, because it can be configured with traffic that models real flows in your networks, across all technologies and with real subscriber data. This is a crucial differentiator, as it helps operators emulate what’s really happening in their networks. So, what kind of data matters? Well, key statistical indicators include:

• Data usage per subscriber (time, day, month, for example)
• Time of day for the consumption of different services, to determine peak hour for the whole subscriber base
• Device and service behaviour for the most popular services (web browsing, video & Music streaming, gaming, IP packet sizes, etc)
• Number of calls and messages per subscriber and time slot (time, day, month)
• Length of call and data in messages (file, video sharing etc)

Understanding traffic flows is the first step and this information can be collected from network logs and records. But, traffic is dynamic – it ebbs and flows, so while snapshots are essential, we also need to take into account changes and growth. For this, we must also factor in:

• Increase of number of subscribers and devices
• Increase in data usage per subscriber (increased usage of popular services and new services)
• Future device and service behaviour for most popular services (web browsing, video & Music streaming, gaming, etc)

This information can also be estimated based on historic performance, allowing a model to be created and configured in Solver that reflects the actual situation in the network and also allows future behaviour to be explored. This is particularly important when considering, not only the existing service mix, but also the impact of adding a new service. What will happen to this through time, given the background of traffic flows and changing demand?

So, a key step in any test programme is to configure Solver with statistical profiles for traffic type, volume, periodicity and more. This ensures adaptation to the unique situation of each operator’s network, through a preparation phase which can easily be accomplished, working with Emblasoft or an integration partner. Of course, there are out-of-the-box capabilities for a quick start, but the long term value comes from adopting real traffic profiles from each operator.

Testing isn’t a one-off activity that’s performed periodically or when a fault is found. It’s a continuous effort that is part of maintaining the health of the network and preventing issues from impacting subscribers. It allows you to discover problems before they disrupt experiences, protecting your reputation.

As such, it simply has to reflect what’s really happening, so that deviations and performance can be explored against a background of changes and loads. The agile approach enabled by Solver allows operators to ensure that their unique situation is reflected in tests and simulations.

The Solver approach is proven in networks around the world. We know this methodology works and we know the value of using real situations to model and emulate traffic. If you want to know more about how to incorporate Solver into your network and to see how Solver is uniquely adapted to your situation, why not get in touch?