Technical Blog

All the latest technical and engineering news from the world of Guavus

Intertwining of Data in 5G Fabric Core

By Saikat Prabhakar, Senior Manager Engineering, Guavus, Inc.

The Fifth Generation

It has not been that long since successful 4G rollout stories were the order of the day; and now the world is gearing up for the Fifth Generation aka 5G. The much awaited wireless standard will revolutionise technology as we know it, in a disruptive way. 5G networks shall enable a diverse range of industries and applications by providing extremely high bandwidth and ultra low latency, which is currently impossible using the existing network standards.

Gartner predicts that by 2020, 3% of CSPs will launch 5G networks commercially and that it will be a key platform enabling technology for artificial intelligence everywhere. With its massive MIMO backed RF capabilities, 5G is going to redefine the way we perceive communication, which is increasingly going to be machine-to-machine.

A Myriad of Technologies

5G is not just a terminology, but rather an ecosystem. An ecosystem that will thrive on essential architectural concepts namely SDN, NFV and MEC among others. While SDN and NFV do not depend on each other, their convergence is inevitable during the course of 5G deployments. SDN helps decouple control plane and commoditise data plane. Whereas, NFV uses defined frameworks like MANO to decouple functionality from appliance and create VNFs across the network and cloud. Current industry use cases such as SD-WAN and vCPE are signifying their coherent emergence. With such an agenda, data analytics can not be an afterthought. In fact, it will play a major role to meet the QoS and QoE requirements and orchestrate the virtualised NFs.

Furthermore, with 5G rollouts, IoT/IIoT is going to take the centerstage and will create vast data lakes every second. This ever evolving holy grail of data would be the prized asset that every Service Provider would love to leverage for profiting on Managed Services. But given the volume and velocity, CSPs will need to quickly invest in it and not just hoard it. The AT&T’s and the Ericsson’s of the world have already ventured in the analytics space. At the same time, data analytics vendors with proven telecom expertise such as Guavus are working towards bridging this gap between investment and profitability. Guavus is investing considerably into developing a strong arsenal of tools, services, and practices to approach the CSPs not just as suppliers but as potential partners.

Data Percolation in the 5G Fabric

Data percolation is inevitable throughout the 5G fabric, therefore analytics vendors will need clearly defined strategies, most importantly, cleaning and wrangling of the vast amount of ‘unstructured’ data being generated (emphasis on unstructured!). The expected result is accurate data models that meet precise standards and business rules as they integrate with other open interfaces.

Another considerable requirement will be support for high speed data in motion. Vendors with proven expertise in real time data processing support and analytics will come out as clear winners. Fast data will be the need of the hour for CSPs as it will help them to mobilise services quickly and deploy new IoT use cases. Capabilities of OSS/BSS providers will also augment with real-time decision making capabilities. Moreover, most of the essential automated network control as expected in a 5G core, can only be achieved with a robust real-time big data strategy in place.

The other major requirement is security compliance and privacy. Most often, the data that gets turned into actionable insights is transactional in nature, which means it includes social and financial information – none of which can we afford to have leaked. Besides, with ever increasing sophistication of threats and cloud utilisation, data governance policies such as GDPR are an absolute must.

The analytics Swiss Army knife will increasingly need to be ML driven that means analytics would evolve from being predictive to prescriptive and finally being cognitive. This will enable network cores to transform into self-healing from being self-learning to ultimately being self-driven.

With the right focus, vendors who are big data specialist will need to partner as DaaS and AaaS providers. These vendors will have to keep a keen eye on complete 5G network life cycle, which shall encompass services created to provide both network as well as application intelligence.

Abbreviations Used

AaaS– Analytics as a Service
CSP– Communication Service Provider
DaaS– Data as a Service
IoT/IIoT– Internet of Things/Industrial IoT
MANO– Management & Orchestration ( NFV )
MEC– Multi-access Edge Computing
MIMO– Multiple Input Multiple Output
ML– Machine Learning
NFV– Network Function Virtualisation
OSS/BSS– Operation Support System / Business Support System
QoE– Quality of Experience
QoS– Quality of Service
RF– Radio Frequency
SDN– Software Defined Networking
SD-WAN– Software Defined Wide Area Networking
VNF– Virtualised Network Function
vCPE– Virtual Customer Premises Equipment


Image attribution:

Posted by guavus