5G is fuelling the growth of ‘Testing as a Service’, as operators prepare to launch new offerings with more enticing service levels.
The arrival of fifth generation mobile networks (5G) is much more than an upgrade to the current standards, as it unlocks a whole host of new use cases and potential business opportunities for operators and customers alike. With 5G, new differentiated traffic classes such as Ultra-Reliable Low-Latency Communications (URLLC) and Massive Machine-Type Communications (mMTC) are now becoming possible – while extending out the customer base beyond consumers to fixed wireless access, enterprise deployments and public infrastructure.
5G will welcome many new types of customers and a whole new class of applications. From remote surgery to augmented reality, 5G will enable applications that could not be delivered with adequate quality on the existing 4G infrastructure. However, the arrival of the next generation of 5G services brings with it several operational and business challenges.
For example, how can operators charge different amounts for different classes of traffic if they can’t measure and prove that the traffic meets the expected quality? And when services are up and running, how do operators ensure that the underlying infrastructure really is meeting the quality requirements of critical applications? This is vital as some of the services that 5G will empower will be life-critical in areas such as telemedicine, autonomous vehicles and emergency response.
To answer both questions, operators must define Service Level Agreements (SLAs) that specify the requirements of the traffic class, the enterprise customer, or the mission critical application and then measure it using ongoing end-to-end testing.
Moreover, testing is more complex within the emerging world of 5G, as operators embrace a wider, disaggregated supplier ecosystem that includes many more software elements running on scalable private clouds. Today, a typical 5G network may have dozens of discrete suppliers across hardware and software elements, leading to a more continuous patching and upgrade cycle where one critical issue can take down an entire service – potentially impacting thousands or even millions of users.
Take the example of building an SLA for a cloud gaming or augmented reality application built to run on a 5G network. The first step is to define the measurement criteria. This starts with defining specific user experience criteria to be measured with a base line of what a “great” gaming experience delivers in terms of real-world performance. Second, the SLA needs to consider geographic reach, defining coverage areas for which an operator will guarantee the quality of service such as a specific venue, city or potentially nationwide. And finally, the SLA must define the conditions under which the agreed level is provided, whether that is on a certain day or between certain timeslots to synchronise with a particular event.
Typically, it takes several steps to create a test cycle suitable to ensure SLA compliance. This starts with defining the test methodology to meet the use case. Next comes the selection of test equipment and the planning and logistics to perform tests. This can include a variety of scenarios including drive test, walk tests and stationary permanent test sets. Next, data from these test platforms are collected within a central repository where they undergo analytics and processing to extract insights. This process can help to both define the initial SLA and when done continuously, can prove an operator is meeting the agreed terms.
Testing as a Service
Even with the deep resource that mobile operators maintain, building a test environment suitable for such a complex environment, and staffing it with people who have subject matter expertise across multiple technologies and vendors, is still a major challenge.
Instead, many are turning to qualified vendor-neutral partners that have dedicated test expertise, products and resource management capabilities as part of a service offering that delivers testing to seamlessly integrate these with other operator functions. This has resulted in significant growth of Testing as a Service (TaaS), both for end-to-end network benchmark testing as in the example above, and with the continuous testing and integration required to make all the disaggregated 5G components work as a whole.
TaaS outsources the testing activities associated with part of an organisation’s critical business activities to a trusted service provider rather than in-house staffing. It is particularly valuable where there is a great deal of change taking place, complexity within the service delivery infrastructure and where any failure will have a major impact to the business and its customers.
But what does this look like in the real world? Some of the first use cases for TaaS are being utilised in the roll-out of the core 5G networks and as an example include a major telecoms operator that has decided to embrace a completely new approach to developing networks and services. Managing a cloud platform consisting of cloud infrastructure and software-based cloud native network functions (CNFs) required the operator to behave more like a software company, a major change from its role overseeing a network of monolithic elements.
This flexibility requires that new CNFs are validated both in isolation and as part of an end-to-end cloud network. To enable rapid collaboration and development, CNF validation must be automated and seamlessly integrated into systems for tracking feature requests, creating new builds and managing the status of testing and bug fixes. This Continuous Integration/Continuous Development (CI/CD) approach is considered a best practice for accelerating cloud software releases.
By using a TaaS, the operator accelerated the deployment of new 5G services. Where previously in its 4G environment, it required work around a four-month cycle of waterfall releases, the use of TaaS helped it shift to agile six-week releases that enables it to adapt to changing market demands much faster. In the first year alone, the operator delivered more than one hundred fully automated validation tests to date with hundreds more planned, reducing the need for additional 5G resources. The shift also helped the operator to overcome key skill shortages by using the TaaS to provide the expertise needed to create a pipeline for rapidly building, testing and deploying new 5G core functions.
Build and evolve
To win with 5G, operators must offer differentiated services and deliver against the expected SLAs. Selecting the right TaaS provider streamlines this process and helps operators bring new products to market that can capitalise on the 5G opportunity. The selection of TaaS should consider several factors, including deep expertise with 5G and the applications running on top of the network. This must be backed by proven testing expertise and proven real-world TaaS deployments. Finally, the TaaS vendor must be a neutral party to ensure bias is not introduced into tests as the number of vendors within the network continues to grow.