Is your IoT Smart?

While it remains to be seen if IoT transpires into the next Industrial Revolution, it has definitely caught the attention of global academics, policy-makers and executives.

While a lot of attention has been paid to the “end device/ sensor” (e.g. smart kitchen, smart home) and how we will be inundated by them (6 per person by 2020, in case you were wondering), less focus has gone into the “intelligence” needed to provide a service. After all having your quota of six devices being merely connected to each other and the Internet hardly delivers much value! What matters is what the interconnectedness does – what intelligent service it provides!

One of the fundamental questions facing the delivery of any intelligent service, is where to embed the intelligence – “smart” i.e. distributed intelligence in the device vs. “dumb” (device) i.e. centralized intelligence. Will the “devices/sensors” be mere data collection and transmission points, or will they be able to (partially or fully) process the data. Why does this matter?

For one, where the intelligence lies impacts the kind of services that can be provided – if data has to be collected, transmitted and processed centrally before the provision of a service, the bias is toward providing less time critical services (e.g. rapid manufacturing), and vice versa.

The smart vs dumb also impacts how the ecosystem evolves. Intelligent devices are likely to drive innovation at the end point (e.g. device vendors win) vs. intelligence at the core (e.g. analytics/solutions providers and telcos likely to win)

Three key factors will drive the evolution of the smart vs dumb ecosystem:
1. Latency – the time required to make decisions. For services that don’t require a near instantaneous response (e.g. Nest’s ability to learn your preferred temperature can wait a few seconds), it is likely that devices will remain “dumb” (the more cost effective option)
2. Privacy – where data transmission results in privacy concerns, it is likely that the device will retain the bulk of the data and process it locally
3. Connectivity quality and cost – the quality, speed and cost of mobile access will drive the balance between transmitting high volumes of data (dumb IoT) vs. local processing with limited inter device transmission (e.g. smart IoT).