5G architecture is so much more than just a new radio technology. It is expected to support new, and as of yet undiscovered use cases, along with new business models; 5G is the edge that network operators really need in order to compete with web scale companies offering their content and services over the top of the same operators’ legacy networks.
There is no doubt that virtualization is going to underpin the whole foundation of these emerging 5G networks. In fact, software-oriented technologies like NFV and SDN play fundamental roles when implementing networks built on 5G architecture.
With the vast majority of underlying infrastructure resources and the surrounding support systems becoming virtualized in 5G networks, it is possible to programmatically package or slice out a complete instance of the entire network. This enables network operators to bring Network-as-a-Service offerings to the market where operators can tailor and optimize packages or slices individually to meet the requirements from completely separate industry verticals such as automotive (a huge topic of conversation at 5G World and Connected Cars in London a couple of weeks ago), healthcare, finance and logistics. This concept of slicing is very different from 4G networks that were primarily designed for one single use case, namely high speed mobile data.
The requirements for 5G vary dramatically across the various industry verticals:
- Massive machine-to-machine communication, also referred to as Internet of Things (IoT) or Internet of Everything (IoE), puts new requirements on scalability, where the 5G network has to support up to 100 times more connected devices than today.
- Critical applications like remote surgery or connected cars will require extremely low latency on the order of single digit milliseconds, as well as constant and seamless service continuity. High availability for these services is also extremely critical.
- Extreme throughput rates. With improved radio technologies, as well as support for optical fixed line access, the throughput rates to be supported by aggregation networks increase exponentially.
A frequently asked question that we hear is: How is service assurance going to be different in 5G networks? We believe that telecommunications can learn a lot from the software industry as it is transitioning from yesterday’s hardware-oriented network infrastructure to tomorrow’s emerging networks, completely defined by software.
In the software industry, there is an established practice to build test cases while developing the target code. This is referred to as test-driven development. The analogy in telecoms would be to construct service activation tests while developing the service specifications and service models- a practice that is not yet extensively done in the industry, but one that is a necessity and that can in fact be implemented already today. This ensures that services are fulfilled correctly and that automated end-to-end activation tests can be performed to confirm that services are delivered right the first time. Also, service models should contain information regarding which KPIs matter most for the customer so that orchestrators can programmatically interpret the model and configure the assurance platform accordingly, in a hands-off, automated way.
Regarding the assurance platform — one thing is clear — it will no longer be an option to use physical test and measurement tools as there will be no physical ports to plug these tools into in a virtualized environment! Instead, assurance solutions must be based on virtualized test agents that can be programmatically deployed according to the needs of designated network slices and dynamic service chains. As each individual slice is designed and tailored for specific requirements, the assurance platform’s test agents must be fully configurable to actively measure those vital KPIs for each individual network slice or service chain, especially when it comes to those key requirements for 5G:
- End to end latency
- Network throughput
- Service and application availability and performance
In order to offer Network-as-a-Service, it is also preferred that the support systems, including the assurance platform, provide multi-tenancy capabilities in order to avoid instantiating a multitude of separate systems, which drastically increases management overhead and operational costs and resource requirements.
Making the transition from current hardware-based networks to programmable and software-defined counterparts is surely a challenging task. However, this is a transformation that all network operators will experience. The sooner this journey is pursued, the more time there is to experiment, learn and improve.
On a final note, one crucial aspect that cannot be overlooked when undergoing this transformation is customer experience. At Netrounds, we provide a comprehensive, programmable test and service assurance platform that provides insight into physical, hybrid and virtual networks, as well as network slices and dynamic service chains, while also allowing visibility into individual customer quality of experience.
So why not start taking your first steps towards an assured network that supports IoT, network slices and dynamic service chains? Please contact us at Netrounds to find out more about how we can help you to assure your next generation network architecture.
Mats Nordlund, CEO & Co-founder, Netrounds
5G World Summit, London