IoT gets redefined…again

At one time, IoT was all about connecting to the cloud. And the widely held belief was that by connecting these things, we will get new information such as the location of the smart thing, health, and other environmental parameters, functional data etc. And all of this data was going to be analysed in the cloud and processed so that new experiences, insights, optimisations and actions could be defined, detected, delivered.

And the belief was based on two premises:

1. The Things were either not smart or were too expensive to run advanced functions beyond keep alive or control. In addition, in the case of mission critical things such as controllers or industrial robots, the smart things were better off dedicating themselves to core functions without letting any other workload affect their main application.
2. The cloud had more compute power and therefore would be better, faster, cheaper in consuming huge amounts of data and converting into actionable information.

So what happened? Hundreds of IoT platforms popped up, offered by large companies and startups alike, each one claiming to have more capability or ease of use or lower cost than all the other similar sounding competitors. And it was widely believed that connecting the thing to the cloud would bring all the benefits attributed to IoT.

But in the last two years, IoT is getting redefined because the beliefs have been disrupted by the following three things:

1. Non-pervasiveness of cloud in Industrial IoT (IIoT): The belief that cloud connectivity is needed to unlock the value of IoT was questioned when it was discovered that from all the IoT segments such as consumer, home automation, IIoT had the biggest unlocked value and therefore the highest ROI for investments. And the IIoT is focused on factories, refineries, and electrical grids where the internet and cloud connectivity is not easily available today.
2. Workload consolidation in smart things: The industrial things such as robots and large machines have sizeable compute capability already built in and will have even more capabilities in the near future. With the development and adoption of workload consolidation, the machines are able to perform multiple tasks with a varying level of mission-criticality and flexibility.
3. Distributed identity through blockchain: With the development of blockchain and distributed ledger technologies, the need for dependence on the cloud for security and authentication is also decreasing. In the future, the ‘things’ will be able to authenticate, transact and interact within the confines of a refinery, factory or a hospital without having to go to the cloud or an on-prem data centre for everything.

And because of these three trends, in particular for the IIoT, the role of the cloud, or even an on-prem data-centre in IoT is beginning to get limited.

The future of IoT is smart things running multiple workloads communicating with each other in real time through blockchain or other distributed ledger technologies with limited dependence on a cloud capability.

It is useful to go deeper into the trend of workload consolidation that is changing how we view Internet of Things and edge. So what is workload consolidation? It means that the new software infrastructure for mission-critical devices enables the running of multiple workloads on a single system or machine without compromising the mission-critical capability of the control function. With the sufficient compute and appropriate software capability, a machine can not only run the controller reliably, but also run workloads for analytics, blockchain, user interfaces or IoT connectivity solutions.

By adding blockchain to this as one of the workloads, the IIoT enabled machines are able to reduce their dependence on cloud or another central authority. The machines can now communicate with each other, authenticate themselves, suggest changes to their scheduled tasks and get them approved, all in real-time.

This actually started way back with autonomous cars. In the case of cars, the two terms were called V2V (Vehicle to Vehicle) and V2I (Vehicle to Infrastructure). The premise was that most of the autonomous driving can happen by leveraging the capability of the car itself or by the car talking to other cars, a.k.a V2V (Vehicle to Vehicle). The whole system was built on the premise that there is either no access or at best an intermittent access to the cloud.

The same is happening in the rest of the IIoT. Most of the work, analytics, and optimisation will happen in real-time at the machine level. That means that the edge and things become really effective and do much more than they have ever done before. In the future, the things in the IIoT will have significant compute capability and will be developed with future proofing in mind, so that the security functions, analytical capabilities, optimisation functions and the control algorithms can be upgraded without a need to change the machine itself.

So what does that look like? Let’s take a few examples:

1. Machine controller – Today a machine controller is typically in place for about 10-15 years and is kept isolated from all the other systems. In future, the control workload will be kept partitioned off, but the controller will have advanced capabilities such as computer vision, event engine, firewall, and third-party apps.
2. Healthcare monitoring machine – Some of the large critical machines such as respiratory and anesthesia care, today take on basic functionality and the doctor is responsible for all the decision making. In future, the doctors will get much more help with smart intelligence on these machines that can analyse data, map data types, and identify trends to help accelerate the decision making process.
3. Transportation – As Hyperloop and high-speed rail functionalities get developed, they will need intelligence along the track. And these intelligent nodes will work with each other to execute the system with low latency and high reliability.

As the smart things and the edge take on the intelligence functions, some of the functions, such as below, will always be done in the cloud such as:

1. Secure onboarding – Bitcoin was an example of software that had onboard incorporated into it with the software provided. But in the industrial environment, the call home and provisioning function is important
2. Software Updates- Software updates will still need to be done. This is also referred to OTA (over-the-air)
3. Maintenance – Any sort of maintenance, especially security patches, will still need to be done, since this is usually provided by a software vendor
4. Workload adjustments – As workload consolidation becomes more prevalent, the workloads might be adjusted to use different types of analytics or new capabilities.

The trend of IoT and digital transformation continues, but in the future, the smart things will play a much bigger role than the cloud especially in the IIoT.

Courtesy of Wind River.