How 5G network slicing redefines telco data analytics strategy
Since 2017, when The Economist published a short article on data being the new oil of the 21st Century, industries across the globe have vigorously turned their focus to procuring, processing and analysing data.
Over time, like crude oil, they have realised that in order to derive the full potential of data, it needs to be refined and processed. Data relevant to procurement, operations, consumption, context of consumption etc., is being leveraged by many of the new age businesses. This, says Vinay Devadatta, practice head innovation & industry relations at Wipro Ltd., has resulted in high levels of ecosystem efficiency and customer satisfaction, that was not considered remotely possible just a decade back.
Digging deeper in data wells
As the world digs deeper into the data wells of derived and contextual data, and explores new data wells (data sources), organisations have started to realise that data can be overwhelming and so, have started drawing boundaries. On top of this, governments have started passing privacy laws, like GDPR (the EU’s General Data Protection Regulations), on collection and leverage of data; hence the focus is now moving towards processing of information.
While data is key the ingredient, effective exploitation requires a well thought out analysis of the same. Industries and organisations have started pursuing specific data analytics strategies, which are aligned with the industry needs and organisational goals and objectives.
The communication industry has been putting various types of data it collects to good use and has developed, over time, various data analytics strategies. These strategies have been effective until now due to uniformity of production and consumption of best effort voice and data services.
The advent of 5G, has brought in the possibility of rendering highly differentiated services with the potential to reimagine core business of industry verticals. With 5G, CSPs (communication service providers) can provide highly specialised services with service guarantees (e.g. quality, accessibility, availability, integrity, mobility). However, this comes with potential of inefficiency unless the resources are managed efficiently.
Typical data analytics in the communication industry have focused on diagnostics, predictions and consumption behavioural insights. In the 5G era, all of three functionalities will play a vital role in assuring not only an optimised utilisation of resources at the disposal of CSPs, but also a satisfied customer.
5G associated differentiation and Service Guarantees are dependent on the unique concept of end to end dynamic Network Slicing. This itself builds on new technologies and concepts such as NFV, SDN, edge computing & data centre distribution. With these new concepts and the new dispersed vertical industry market, the role of data analytics becomes even more crucial and strategies and functionalities need to be further honed to specific needs of 5G deployments.
Distributed analytics: Localisation of services is an inherent tendency, and this results in a lot of contextual data being generated, which is pertinent to specific localities. Additionally, as new rounds of privacy rules are expected to kick in, organisations will be forced to limit propagation of data across ever converging geographical boundaries. Recent focus on Mobile Edge computing at the Edge Data Centres is another reflection of this change.
Besides this, some of the 5G related services may require low latency decision making. All of these indicate the need to carry out analytics of communication data right at the edge, in data centres distributed across geographies. Future data analytics solution should be amenable to wide distribution with centralisation of data and analytics limited to compelling reasons only.
Differentiated analytics: 5G brings in the possibility of rendering differentiated services to industry verticals. The new market of vertical businesses also faces the issue of dealing with a wide variety of operational and business models. Besides this, effective use of flexibility brought in through virtualisation also needs specialised real time decision making for optimised deployment.
These require data analytics architectures which can ingest, process and analyse data differently based on nature of data, as well as the target consumption of analysed outputs, e.g. real time, batch processing and using Artificial Intelligence techniques. The system should support multiple parallel channels of analytics.
Drill down support: Unlike many of the “Uber” generation businesses like taxi services, accommodation services etc., there is no direct mapping between the resources used for production of services with the form in which they are consumed. The services of infrastructure level resources are transformed through layers of abstractions (e.g. networking equipment, network service, Network slice, Communication service) before reaching the end customer.
Also due to virtualisation techniques and dynamic slicing, the relations between entities at different layers of abstraction is ephemeral. Since CSPs will be offering service guarantees, in the case of a service breach; it should be possible to drill down to the root cause, based on the transient relation at the time of incidence.
Demand driven: Differentiated services, provided through 5G slicing, require varying types of specific analytics to ensure they adhere to service guarantees. Currently analytics flows bottom up, from resource layer to customer layer. Since it is not possible to carry out all types of analysis on entire data, it should be possible to request for specific analysis to be carried out at any layer of abstraction on specifically identified entities.
Decision support: 5G Network Slicing results in bucketing of resources, which can potentially result in resources contentions within and across slices. As the services are differentiated there can be vast differences in revenues earned based on allocation of resources.
Besides, CSPs also need to strike a balance between immediate revenues and strategic goals. All these decisions require in depth analysis based on ever-changing situations, while taking into consideration predicted future demand. This requires a well-structured what-if analysis, so that automated decisions can be taken.
In conclusion, CSPs need to evaluate and revamp their data analytics strategies and solutions in the light of 5G and the changing consumer market. While procuring data, processing and deriving actions is essential, optimised deployment of 5G Network Slicing requires well thought of implementation of Data Analytics.
The author is Vinay Devadatta, practice head innovation & industry relations, Wipro Ltd
About the author
Vinay has extensive experience in the communications industry, working with equipment vendors, communication service providers, OSS product vendors, OSS solution providers and telco management standards bodies like TM Forum, ETSI NFV, NGMN NGCOR. He is responsible for leveraging external and internal innovation and creating new service lines at Wipro.
His expertise is in the areas of OSS, NFV, Orchestration, automation, digital customer experience and 5G . He has led multiple cross-organisation projects exploring new concepts and futuristic Service Provide needs. He holds a B.Tech in Electronics & Communication Engineering and an M.Tech in Computer Science & Technology from IIT Roorkee.