Programmable Networks Service Ai ML Operation New
Programmable Networks Service Ai ML Operation New
The ‘concept of zero' approach is a growing trend among telcos, implying zero
human-led latency, zero vendor lock-in, zero limitation on the capacity, and zero
defects. It aims to develop a software-based network by leveraging SDN technologies
that deliver centralized control, programmability, virtualization and a high level of
automation combined with artificial intelligence (AI) and machine learning (ML).
These are the guiding principles behind the ‘concept of zero'. Leveraging AI-ML,
telcos can process and analyze huge volumes of data to extract actionable insights
that help enhance customer experience, improve ease of operations of complex
networks, and increase revenue through new products and services.
Autonomous networks enable telcos to provide network-as-a-service. This approach ensures high
degree of flexibility, reconfiguration and enhanced functionalities through software updation,
resulting in improved end-user experience and reduced opex.
Cognitive network operations are crucial in tackling the complexity of managing virtualized
networks across radio access networks (edge computing, virtual RAN, open RAN), core (cloud
native, enterprise SD-WAN), transport (SDN), and next-gen OSS.
The big leap from automation to
autonomous networking
Autonomous networks have been enabling digital transformation across organizations. An
autonomous network represents a powerful paradigm shift that runs with minimal to no human
intervention and can configure, monitor and maintain itself independently, leveraging a variety of
new-age, disruptive technologies. Autonomous networks use artificial intelligence and cloud
technologies to become self-provisioning, self-diagnosing and self-correcting.
However, there are several challenges that programmable networks face in transitioning from
automated to autonomous networking. These include use cases for the networks that have been
using new technologies such as multi-tenancy and network slicing. Today’s networks add a level of
complexity as they change dynamically, and the ability to maintain an accurate view of the state of
the network in real time is a challenge. In addition, knowing how to manage changing topologies
and assess how they impact management actions (i.e. effective actions) in autonomous networks
is critical. As networks are expected to become more on-demand and therefore highly specialized
in a given context, the operator may not necessarily benefit from a large amount of historic data
as it may be too specific.
The challenge is in deploying cognitive network management and its orchestration across multiple
heterogeneous networks, all of which have their own peculiarities and requirements in
multi-vendor and multi-technology network landscapes. This includes radio and access networks,
core and aggregation, edge networks, edge and computing clouds, and satellite networks.
The approach also enables end-to-end cross domain correlation to deliver autonomous decision
making - complex decisions based on the detection of a large number of hidden or hierarchical
influencers and self-healing zero touch operations.
The ‘zero concept’ approach targets the development of software-based network by leveraging
SDN technologies that deliver centralized control, programmability, virtualization and an extreme
level of automation combined with AI-ML concepts. These are the guiding principles behind the
‘zero concept’, an in-demand trend from telcos and other enterprises. Broadly, it means zero
human-led latency, zero vendor lock-in, zero limitation on the capacity, and zero defects.
This journey toward the ‘concept of zero’ is aligned to the network infrastructure demand that
needs to closely work with network functions virtualization infrastructure (NFVI) and NFV
orchestrator (NFVO) integration requirements and the maturity in the SDN technology and its
market adoption.
The 'concept of zero' approach has its origins in the traditional legacy networks, which involves
the three stages of transformation – automated to adaptive to the final leg of the transformative
journey, autonomous.
Automated
The ability of the network to adapt to changing business requirements depends on the extent
of automation in operations and the implementation of network changes. Currently, telcos
have dedicated teams to manually schedule the device back-up jobs, change policies on a
multitude of devices, perform provisioning with command-line interfaces (CLIs), and spend
hours in troubleshooting and restoration.
The philosophy behind automation is to introduce a basic level of programmability, centralized
control and virtualization. This is the first step toward transitioning to autonomous networks.
Adaptive
Beyond automation, enterprises are now able to re-think their network operations with
advanced analytics to achieve their long-desired goal of being adaptive. For this, networks
need to be aware of the applications, transactions, and alternatives available for resolution
along with the required actions.
Data center interconnect demands are changing rapidly with cloud computing, virtualization
and growth in traffic. Utilization of the resources spread across various data centers can be
drastically improved by pooling and sharing. Bandwidth-on-demand increases network agility,
leading to building business responsive networks.
Autonomous
An autonomous network requires zero to minimal human intervention. It brings in intelligence
from ‘closed-loop network telemetry’ and provides the business context. Artificial intelligence
and machine learning techniques are employed in order to make the network self-driving and
self-healing.
SDN attributes (drive, differentiate, digitize, migrate and intelligent branch network cloud) are
aligned to the network journey from traditional to automation to adaptive to autonomous, the
end goal of which is the ‘concept of zero’. Adding to this transformation is technology maturity.
As part of the 'concept of zero' transformation, service assurance is centered on collecting,
correlating, and resolving events in an automated, predictable, machine-first approach to
address the following aspects:
Maximize customer experience through proactive identification and resolution of issues
Minimize operational expenditure by maximizing automation
Strive toward unmanned/lights out NOC through big data analytics, machine learning and
robotic process automation (RPA)
Minimize customer service expenses through automation-driven approach with proactive
notifications
Enable insights across different types of monitoring information with single touch collection
of all monitoring information (events, performance metrics, traffic information, customer
behavior) through unified data ingestion platform
Leverage non-traditional inputs such as social media feeds, weather information and events
such as sports
Navigating the future of telecom
with AI
AI technologies have the potential to completely transform the telecom sector, the most distinct
applications being the classification of traffic, anomaly detection and network optimization.
In addition to these applications, AI helps in autonomous decision making, self-optimization,
self-learning and self-healing.
With the growing endorsement of network abstraction technologies such as 5G, edge computing,
SD-WAN and NFV, improved network visibility and effective network management can be
achieved with AI-ML based programmable network operations. This will facilitate autonomous
decision making toward future-proof and self-heal network operations.
Enhancing the telecom industry with AI could lead to various new revenue streams. The AI-ML
based programmable network operations have the ability to address multi-vendor complexity to
achieve end-to-end automation across domains. Not only will this improve customer experience
with a single pane of view for the wireless, wireline and enterprise networks, but it will also allow
for opex optimization.
About the authors
Tanuj Duhan
Tanuj Duhan is a senior domain consultant with TCS'
Communications, Media and Information Services business
group. He has 20 years of experience in telecom, digital and IoT
domains. Tanuj is a key contributor for delivering cognitive
solutions to telcos, helping in their automation journey across
wireless, fixed line, cable and enterprise domains.
Lovaljeet Kaur
Lovaljeet Kaur is a technology consultant in TCS'
Communications, Media and Information Services business
group. She has 11 years of experience in the fields of AI analytics
(NLP and Computer Vision), automation and data networks.
Lovaljeet is an avid contributor to the open-source community
and specializes in delivering potential AI analytics use cases.
Contact
Email: [email protected]
TCS offers an integrated portfolio of cognitive powered business, technology, and engineering
services and solutions. The company's 469,000 consultants in 46 countries help empower
individuals, enterprises, and societies to build on belief.
Visit www.tcs.com and follow TCS news @TCS_News.
All content / information present here is the exclusive property of Tata Consultancy Services Limited (TCS). The content / information contained
here is correct at the time of publishing. No material from here may be copied, modified, reproduced, republished, uploaded, transmitted, posted
or distributed in any form without prior written permission from TCS. Unauthorized use of the content / information appearing here may violate
copyright, trademark and other applicable laws, and could result in criminal or civil penalties.
Copyright © 2021 Tata Consultancy Services Limited