Living on the Edge: the power of Edge computing in public service delivery

By on 26/06/2019 | Updated on 26/06/2019

‘Edge’ computing is the new kid on the AI block, and could dramatically expand the public sector applications of machine learning; but it demands a specific set of skills and capabilities. Marcelle von Wendland explores the potential and requirements of this emerging technology

Artificial Intelligence (AI) and machine learning are moving out of the cloud and into the sharp end of public service delivery, as the emerging technology of ‘Edge’ computing gathers momentum. But to realise the opportunities – and avoid the risks – of this new form of AI, public servants will need the human intelligence and the organic learning to commission and manage it. For, like any new technology, Edge has unique characteristics and demands specialist expertise.

AI allows systems to adapt to their individual users and local environments, ‘learning’ their specific needs and behaviours – and thus improving decision-making. That requires enormous computing power, a requirement that – until recently – confined its operation to the cloud, where vast data storage and processing capacity is available.

Expanding the boundaries of AI

But this constraint limits its applications. In locations without fibre-optic connections, system managers must depend on technologies such as 5G to transmit data to the cloud – and given the volumes of data demanded by many AI applications, this is often unfeasible or uneconomic.

In some cases, speed is so critical – in emergency warning systems, for example – that there’s no time to pass data back to the cloud and await a response.

And sometimes, uploading data to the cloud might compromise citizens’ privacy or data security. By using local sensors to monitor the movement of smartphones or vehicles, for example, public servants can gather the real-time information to improve crowd control, traffic management or air quality. But transmitting this raw data to the cloud would risk security breaches, and could permit individual members of the public to be identified without their permission.

The unique potential of Edge

One solution is Edge: by locating the processing power close to the sensors gathering data, at the system’s edge, public servants can realise the benefits of machine learning tech without paying a price in cost, time or security. Today we are developing computing systems that are compact, hardy and energy-efficient enough to operate at the frontline of service delivery. And by anonymising data before it leaves the Edge, we can deploy them without compromising people’s privacy.

As an illustration, we at Fincore – a software business whose roots lie in risk management and financial services – are working with a railway operator to install systems that will detect trespassers on the line.

Differentiating people from animals, flying leaves or other detritus demands machine learning technology, whilst rapid decision-making is safety-critical; and in many areas we’re dependent on 3G connections. So the cloud isn’t an option: by installing a range of sensors and the required processing power trackside, we can perform this essential function using Edge – cutting the need for staff to walk the line, whilst improving detection of trespassers.

To take another example, in urban areas Edge AI systems can learn local road use, wind and pollution patterns, building detailed models that enable public servants to dynamically manage traffic and parking. By containing the raw data at the Edge, we can alleviate congestion and improve air quality without compromising people’s privacy or data security.

Mastering the technology

To make use of Edge’s potential, though, public servants must address a unique set of technological and managerial challenges.

Edge equipment must be resilient enough to survive poor weather, for example, and to resist theft – protecting both its components, and the data they contain. It must operate autonomously and reliably in all conditions, bringing heavyweight processing power out of its usual habitat: carefully climate-controlled server centres, with maintenance and support staff on site. And it must often generate its own power using renewable technologies, enabling its use in off-grid locations.

So the staff commissioning Edge systems need sufficient expertise in its specific characteristics and requirements, as well as a wider knowledge of AI and machine learning technologies. They must be able to assess providers’ own technological prowess, and ensure vendors have the skills to understand and address the requirements of public sector delivery environments – such as regulatory, legislative, reporting, safety and privacy issues.

The power of partnership

On the management side, Edge systems often involve a range of public and private sector stakeholders: in traffic management, these might include local authorities, highways agencies, utility companies, and an ecosystem of suppliers. So commissioning and operating Edge can demand highly collaborative working, with dedicated arrangements for project management, finance and oversight.

And there is an art to successfully commissioning AI capabilities. Whilst it can seem most straightforward to buy the necessary hardware, it’s often better to specify the outcomes required, agree a set of service level standards, and purchase the capability as a service. This incentivises providers to ensure the system works as expected – but demands a specific set of procurement and contract management skills.

By bringing computing power to the point where it’s required, Edge computing can get around many of the restrictions on the use of AI and machine learning – enabling public servants to address previously insoluble problems. But it shouldn’t be viewed simply as ‘cloud 2.0’: Edge is a new branch of computing, with its own very specific characteristics and requirements. And if your problem demands a solution this bespoke, your means of commissioning, delivery and management will need to be just as unique.

Edge promises to expand the power of AI into whole new fields of public service delivery. To realise its potential, public servants will need new skills, systems and partnerships. But if they can master its unique demands, the rewards could be enormous – improving efficiency, improving outcomes and, ultimately, providing better services for the public. 

Marcelle von Wendland is a Consultant for technology and software provider Fincore, and the head of Finworks’ data business. For more information, visit To get in touch, email her at [email protected]

About Partner Content

This content is brought to you by a Global Government Forum, Knowledge Partner.

One Comment

  1. Elaine Bergeron says:

    It sounds like Edge computing is worth considering as an alternative to the growing needs for “data lakes” and supporting infrastructure.

Leave a Reply

Your email address will not be published. Required fields are marked *