- Network providers anticipate that tens of billions of devices will be connected during the next 10-15 years, fuelling an explosion of new services and placing the current network infrastructure under unprecedented pressure.
- The collaborative multi-disciplinary research programme, Next Generation Converged Digital Infrastructure (NG-CDI), is creating capabilities that will enable the UK’s telecommunications and digital infrastructure to respond to these pressures.
- Led by Principal Investigator Professor Nicholas Race from Lancaster University and Stephen Cassidy from BT, NG-CDI addresses the challenges posed: how to manage a complex infrastructure with increased response speed while also simplifying its operation and dealing with future developments, such as 5G and self-driving vehicles.
- Research outcomes have already meant that BT can introduce an array of innovative management processes, capabilities, and decision-support tools.
- NG-CDI and BT developers are building the essential foundations to deliver future telecommunications infrastructure.
Since the introduction of the Strowger telephone exchange in the late 19th century, networks have become progressively more automated. Automation has enabled humanity to transfer our efforts to higher-level activities such as engineering design and proposition development – including the research process that produces ideas and gauges marketplace reactions for new or improved services or products.
Fast forward to today, and network providers such as BT are confronted with the challenge of how to increase automation to meet growing customer demands, as well as those of the broader telecommunications ecosystem. These requirements are propelling an eruption of complexity and the agility needed for operations and service delivery. Network providers expect a massive growth in new applications and services to be used in almost every area of our professional and personal lives – and anticipate that tens of billions of devices will be connected during the next 10-15 years, placing the underlying infrastructure under unprecedented pressure.
Principal Investigator Professor Nicholas Race (Lancaster University) and Stephen Cassidy (BT Research) are leading the Next Generation Converged Digital Infrastructure (NG-CDI) in the UK, an ambitious collaborative programme comprising a multi-disciplinary team of academic researchers from the University of Bristol, the University of Cambridge, Lancaster University, the University of Surrey, along with BT’s industry researchers. Together, they are creating the next-generation management capabilities for the UK’s telecommunications and digital internet infrastructure. Supported by the Engineering and Physics Sciences Research Council (EPSRC) ‘Prosperity Partnership’ scheme, the project fosters strategic, research-based partnerships between academia and industry, co-investing and sharing research challenges.
Network providers such as BT are confronted with the challenge of how to increase automation to meet growing demands.
Vision for an autonomic infrastructure
NG-CDI is addressing the challenges posed by the complexity and scale intrinsic to future developments such as 5G and self-driving vehicles that will require a fundamental shift in the performance, maintenance, and management processes of our networks. Advances in programmable data-driven networking underpin the project’s vision for an ‘autonomic’ infrastructure that can both predict and react to changes in demand. This will allow telecommunications equipment to be suitably reconfigured with minimal human intervention. The researchers explain how this agility is central for new services and providing better customer experiences, as well as improving network resilience and operational efficiency.
Telecommunications is recognised as an essential part of the UK’s Critical National Infrastructure, enabling the UK to continue to prosper as a leading digital economy. Realising the NG-CDI’s vision requires a wealth of experience in the operation of national and international digital infrastructures. In addition to this profound understanding, world-class expertise is needed in the fields of data analytics, networked systems, machine learning, network functions virtualisation, cyber-physical systems, business innovation, and asset management.
Net DevOps approach
As network infrastructures become increasingly software-based, Net DevOps (Network Development Operations) offers workflow automation opportunities that would have been impossible with a hardware-based infrastructure. Inspired by cloud software DevOps, NG-CDI brings together people, processes, and technologies to increase automation and reduce deployment times. Networks pose additional challenges, such as heterogeneity and scale, so technologies must be modified to leverage CI/CD (continuous integration, continuous delivery, and continuous deployment) in network infrastructures.
To understand how a network led by the behaviour of its millions of users would behave, and to safeguard the infrastructure’s performance demands, the collection and careful monitoring of numerous data streams is needed. Both data quantity and diversity will increase with the data-driven, autonomic networks of the future. It requires accurate, timely detection to recognise any anomalous patterns which may suggest operational interventions to minimise service disruption.
NG-CDI has created a suite of computationally efficient anomaly detection methods underpinned by the rigour of statistical inference. These can run in real-time and distinguish, for example, if the raw data is a brief point anomaly or a more persistent collective anomaly when compared with the usual baseline. Race and Cassidy say that BT’s operational systems are already using these vast ‘online anomaly detection algorithms deployed into live BT operational systems to automatically detect anomalous network behaviours requiring human intervention.’
Academic and BT researchers co-work closely using the huge operational data produced by BT’s network and processes.
This project is producing a range of outputs that cover a variety of areas of impact with timescales ranging from immediate to long-term. Academic and BT researchers co-work closely using the huge operational data produced by BT’s network and processes. These solutions are already embedded and can therefore be swiftly integrated into the business. BT researchers bridge the gap between academic research and the appropriate operational and business areas domains. Network testbeds facilitate trials of research outputs in real-world environments.
Research outcomes have already enabled BT to introduce an array of innovative management processes, capabilities, and decision-support tools. Researchers have also created pioneering intent-based interfaces that can support the definition of new services in business, rather than technical language, as well as their subsequent scaled rollout and continuous monitoring. Developments include novel methods for optimising network design together with engineering processes. Models that expand both our understanding and decision-making have also been demonstrated.
The NG-CDI research programme is moulding BT’s strategy and vision for the future. Moreover, it is shaping the wider telecoms industry via its engagement with Standards Organisations. In collaboration with the TM Forum (TMF, a global industry association for telecommunications service providers and their suppliers), NG-CDI and BT developers are building the fundamental foundations to deliver the future telecommunications infrastructure.
What inspired you to conduct this research?
The realisation that to generate the agility needed to deal with future services that are not yet predicted, and with sufficient speed, BT needed a radically new approach: an autonomic, virtualised software-enabled infrastructure.
What are the main barriers to futureproofing our telecommunications infrastructure?
The very large sunk costs and the need to bring the whole industry along (manufacture, operators, standards) to create economic scale at a global level.
Which things have surprised you during this project?
How prescient we were! What was a ‘glint in the eye’ originally has now become what customers (large corporates, service providers) are beginning to demand.
What advice would you give a young researcher interested in getting started in your field?
Be prepared to be multidisciplinary – to fit technical expertise into an evolving context.
What is NG-CDI’s most exciting research outcome to date?
The deep connection with the human world – for example, though the introduction of ‘intents’ that create a continuous ‘conversational’ interface with customers (amazingly valuable), and the integration of roles, tools and motivations/culture which brings together a completely different model of operational and risk management that is transformational for the business.