IoT and the edge – an evolving relationship

The internet of things (IoT) is fast becoming a reality.

The combination of the UK’s developing edge computing platforms and high-speed, high bandwidth 5G and full fibre connectivity brings advanced artificial intelligence (AI) driven applications and services within reach of almost every business in the country. We now live in a world where machines interact with each other more than humans do.

IoT growth depends on edge expansion because it requires a platform with great network connectivity, regionalised compute capability, and cloud access. Without these edge capabilities, businesses will not be able to use IoT applications at scale in the way they already use cloud applications. These need to process the masses of data from the multiple sensors and devices that comprise the IoT.

Shipping all data to be processed at the main public cloud providers’ data centres will become unsustainable due to latency and cost considerations.  To mitigate this, IoT needs gateway hubs to aggregate the data, operate actuators and translate between sensor protocols used to connect to a network. This is best suited to the edge data centre where the gateway will filter out unnecessary data and pass on critical information to proprietary applications hosted in public cloud.

Industry adoption of edge is currently led by early-adopter use of MECs (multi-access edge computing environments) providing IT services, compute, and cloud access. This, however, will soon give way to shared services at the metropolitan level. There are already live use cases within the smart city, transport and energy sectors, but large-scale adoption will only follow once edge infrastructure platforms have fully developed their low latency connectivity, high-speed backhaul to the public cloud and local computing capabilities.

Nevertheless, at enterprise level three challenges commonly stand in the way of IoT adoption, beginning with the need to gain a full understanding of the benefits it can deliver in pure business terms. Then there is the challenge of integrating the multiplicity of IoT devices, gateways, and the data that these generate into an enterprise’s current architecture. The growth of AI applications also means that architectures will have to facilitate more data being transferred back to the edge for decision-making in intelligent IoT systems.

The third challenge is the more long-standing problem of acquiring staff with the requisite skills in data architecture aligned with process business transformation. Organisations can only overcome this endemic difficulty by selecting the right partners with deep expertise in the developing relationship between edge platforms and IoT implementations.

For IoT to accelerate, access to reliable and low latency connectivity is becoming essential. The primary markets for hardware devices will be dwarfed by the market for applications based on continuous streams of sensor data. Applications focused on real-time and aggregated data analytics need connectivity that has either low jitter, loss and lag or has dedicated high bandwidth.

As the IoT develops it will spread across a wide range of business applications that require different kinds of connectivity. Conventional multi-protocol label switching (MPLS) which has been the backbone of the internet for a couple of decades, cannot, however, match the newer iterations of SD-WAN (software-defined networking in a wide area network) for flexibility. But MPLS will continue to underpin many services due to its reliability and scalability.

Unlike traditional WAN architectures which lack the central visibility and control required for distributed IT environments, SD-WAN delivers a step change for businesses. It makes it easy to configure multiple devices, saving time and increasing efficiency. Organisations can enforce their own policy, based on user experience, with network priority given to the most business-critical applications to avoid jitter, lag, or brownouts. Rolling out new applications becomes quicker and less costly across multiple sites.

Although access to edge services does not depend on SD-WAN, the increasing use of software to define and optimise network performance will accelerate the full operationalisation of edge computing platforms and allow an edge value chain ecosystem to fully develop and interact programmatically through APIs.

The growth of edge infrastructure, which now covers approximately 95 per cent of UK businesses, gives organisations the architecture they need for IoT and AI-driven automation, efficiency, and innovation.

About the Author

Simon Michie is CTO at Pulsant. Pulsant is the UK’s premier digital edge infrastructure company providing next-generation cloud, colocation and connectivity services. With a network of 12 strategically located edge data centres, Pulsant brings the advances of edge computing within reach of 95 per cent of the UK population.

Featured image: ©greenbutterfly

more insights