Mobile operators deploying 5G have an entirely new packet core architecture to assemble as they build out their infrastructure; where and how they will get network intelligence in this environment is a key success factor to ensure that they can monetize their 5G networks, as well as maintain high quality as the number of users and use cases grow on new networks.
The 3GPP 5G standard introduced the Network Data Analytics (NWDA) function to mobile network architectures, and has defined a collection of initial use cases that specify what kind of data can be provided to feed 5G network decisions. ETSI has recognized the importance of mobile customer experience and launched an Experiential Network Intelligence (ENI) ISG that leverages artificial intelligence and machine learning on network analytics to adjust services based on changes in user needs, environmental conditions, and business goals. Additionally, ETSI has defined a set of initial use cases that demonstrate how an ENI system would enhance networks. When these two powerful solutions are combined into a single network intelligence solution, they will create a disruptive advantage for a mobile operator – today and in the 5G future.
To a consumer, all that matters is if an application is working, and if it’s working right now. If a YouTube video is low resolution and stalling, Skype voice is skipping, or Facebook is loading images slowly, then the user is having a bad experience. If every other subscriber on the network is getting good service, the network has still failed that one subscriber. The challenge today is that network operators don’t have the visibility to determine that this has occurred, much less to fix the problem. Sandvine’s Active Network Intelligence solutions are served by an unparalleled foundation of contextual data that meets and exceeds the data requirements specified for many of the NWDA and ENI use cases. Using artificial intelligence (AI) and machine learning (ML) techniques, Sandvine takes data collected in an NWDA role and processes it as an ENI engine would, in order to help solve network issues like those described in various NWDA and ENI use cases. The ENI engine then determines the action that needs to be taken, and deploys policies on the network to solve the issues in real-time. This network intelligence enables closed loop automation, since with right data in the right context can enable automation for specific use cases.
What data is needed to determine how the network is performing for users? If an operator can measure throughput, latency, and packet loss for each subscriber, this provides a foundation to understand the experience that the network is delivering to users. If you add context to that measurement (i.e. location, service plan, device, type, etc.), that intelligence becomes actionable to improve the user experience. The chart below gives examples of what kind of metrics are needed by each application type to get a good experience from the network:
If you are interested in understanding more about how mobile analytics will be used to automate 5G networks, download our whitepaper on the topic, and register for our webinar next week. This is the first in a three part series that will explore the principles of NWDA and ENI, and map them to specific use cases that can be deployed today with Sandvine solutions. The first part explores the network data – also known as telemetry – needed to measure experience and how it needs to be contextualized to make it actionable, comparable to an NWDA. The rest of the series will focus on processing and analyzing the data, as an ENI engine would, and turning that into intelligent policies to close the loop – the process that Sandvine calls Active Network Intelligence.