Disruption in Technology
The business world has remained pretty much the same since the industrial revolution — but things are going to change. Disruptions in technology are altering the economics of businesses worldwide. These disruptions have led yesterday’s start-ups to overshadow old establishments in terms of value and impact.
Change in technology outpaces its adoption. The same is true for the Internet of Things (IoT). Offering new products and services requires keeping track of new trends in technology. Disruptions in the IoT landscape are dislodging some of the fundamental and well-established architectural patterns. I discuss some of the significant developments around IoT architectures and infrastructures that are likely to have an impact in the near future.
More Power to Edge
It is not necessary that all devices communicate with the cloud given the pace at which their numbers are growing. The key drivers pushing IoT compute from the cloud to the edge are the quantity of data, network availability, security, and low latency. The edge is going to explode with power of infrastructure. Especially in Industrial Internet of Things (IIoT), the edge software will be running on robust hardware, or even on mini-datacenters, offering data collection, real-time analytics, and powerful machine learning (ML) models.
On the software side, the edge will be more modular in the form of micro services and container images readily available at marketplaces. Moreover, standardization will be achieved in connectivity toward the cloud in terms of security protocols and communication.
Hybrid Integration Platform
IoT solutions will be hugely impacted by the expansion of SaaS offerings. New-age IoT solutions will consume multiple SaaS products from various vendors along with their own applications in data centers on premise. A hybrid integration platform can securely combine on-premise and various cloud-based systems leveraging technologies like TLS. It can bridge the network divide between multiple clouds and the infrastructure on premise, and enables integration of on-premise and cloud end points, allowing organizations to operate with exceptional speed and agility.
Image courtesy: Gartner, Inc.
Specialized Database for Sensor Data
A database to store sensor data is not yet available. Traditional RDBMS or NoSQL doesn’t align with the needs of IoT applications or storing sensor data. The criteria for an IoT storage include scalability, ability to ingest data at high rate, flexible schema, and integration with analytics tools with inbuilt machine learning models. The databases should have specialized data types for storing 3D data, digital signals, device metadata, etc. Time series databases with inbuilt analytics and visualization capabilities have been launched in the past few months, but there still is huge scope for improvement.
I don’t expect speech to replace written communication; however, standard human-machine communication is moving toward spoken interaction and away from screen-based (touch or click type) communication. This trend is more apparent in consumer IoT solutions than in industrial IoT use cases as millennials — the new generation of consumers — are more adept at dealing with gadgets. The mass adoption of voice-first applications is not far away with Amazon and Google launching SaaS API offerings to integrate applications with their Alexa and Google Home products, respectively. These voice-first IoT applications, combined with personalized and contextualized suggestions generated by embedded artificial intelligence (AI), are set to enhance the user experience with machines.
The concept of ‘digital twin’ is not new. All major IoT vendors already support the paradigm, be it Microsoft, AWS, IBM, or GE. Digitally representing a physical object has been around for over 30 years in the form of CAD models, assets models, etc. For example, NASA has run complex simulations of spacecraft for decades.
Gartner predicts that half of all large industrial companies will use digital twins by 2021. We may, however, be only scratching the surface as most digital twins are still implemented as ‘status twins,’ showing the current status and accepting a desired state.
There has been slow progress in cases of ‘operational twin,’ a twin to a process in a plant, and ‘simulator twin,’ a twin to replicate the behavior of an equipment. The advancement in case of compound assets for ‘status twin,’ where a digital twin is composed of several digital twins organized to make a composite twin, has been tepid.
As ‘status twin’ is the foundation for the other twins, progress on these twins is expected in the near future. Once achieved, the combined twin may become ‘cognitive twin,’ which may be working autonomously by making a community with other twins.
These are some of the technology trends relating to IoT that may be relevant this year and in 2019.