IoT: There must be firm foundations under the fog

Suddenly everybody is talking about robots. Has the TV series Humans sparked these discussions or Apple co-founder Steve Wozniak’s claim that if robots take over we will become their pets?

Either way, it all sounds scary. However, experience tells us that most technologies go through a long gestation period between the hype and the reality. For example, the closest we get to machines taking over the world at the moment is the Internet of Things (IoT) – where everyday objects have network connectivity, allowing them to send and receive data.

Hardly a day passes without another prediction of the spectacular growth of IoT applications.

This is a technology which holds less of a threat than robots and certainly huge potential. But still CapGemini estimates that 70 per cent of organisations offering IoT do not yet generate revenues from them. Many are still running pilot schemes with little immediate return on investment. There are also still concerns about complexity and perceived data security issues.

But when it comes down to it, cut through the excitement and speculation surrounding IoT and the challenge lies underneath. The key to unlocking the potential of IoT sits with an organisation’s underlying infrastructure. Unless this is agile, has resilient networking capabilities, effective security and reliable connectivity, a business is as unlikely to bring an IoT application quickly and profitably to market as a robot becoming their next CEO.

The main reason for this is the scale of demands that will be placed on infrastructure by the sheer volumes of data produced. Also the need to connect an increasing number of network endpoints that deliver faster wireless connectively will also become critical as IoT gathers pace.

If predictions of IoT growth are correct, proliferation will have a huge impact on enterprise networks. Many IoT applications will demand ubiquitous connectivity across multiple networks which will need to cope with unpredictable traffic volumes. Anything from a faulty freezer to a natural disaster will cause dramatic peaks as faults and service interruptions are logged and reported. Equally, infrastructure and associated applications will be placed under stress by the need to collect, move and analyse all the data generated.

There will also be the matter of developing interoperable standards because of the heterogeneous nature of the devices and communications protocols in use. Many IT departments will have little experience of the specific technologies in use and risk connectivity and security problems.

However, perhaps the most important issue will be how an organisation adapts and enhances its distributed cloud delivery model in order to cope with escalating traffic and data growth. Today’s cloud architectures could struggle to cope with future IoT networks.

If we get to the point where there are billions of devices in play and vast numbers of data transactions taking place in real time, the idea of a mesh or edge architecture with devices talking directly to each other and handling many of their own computational tasks, could take off. This type of ‘fog computing’ – an environment where data, processing and applications are undertaken by devices around the network edge rather directly in the cloud, is already being heralded as the answer to this challenge, although there are still concerns about security.

Whatever the chosen network configuration or approach, it will be difficult for businesses to go it alone. They will need the security of a technology partner who can deliver networks to the highest standards, manage services and perhaps more importantly provide consultancy and advice on future strategic direction.

So more opportunities for services providers? Certainly, but only if they have the experience and vision to fill the gap between the idea and the practical application.

The IoT will only really reach its tipping point when it is commercially attractive and this can only happen with a firm IT foundation. When the machines take over, however, is a different matter.

Sean Paxton, product manager, Redcentric