Living on the Edge: Can Autonomy at the Edge Drive an Explosion of IoT Services?
IoT, while interesting, is not inherently useful to electronics companies — or actually any company — unless there is a very solid understanding of the outcomes it can deliver and the best way for it to achieve those outcomes.
The existing volume and velocity of IoT data have presented significant organizational challenges. First, there’s a deluge of under-used and unused data that many organizations can’t handle or make sense of, let alone making money off of it. Second, there’s an infrastructure fragility; simply put, our current processing and storage systems weren’t set up for this.
As IoT is at the early stages of its evolution, it’s not fully defined — it’s still morphing into what it will be at maturity. So in order to peek forward we need to place a lens on other technologies that have been experienced a similar trajectory. Here I posit two technology examples that provide proxies for us to examine the currently largely untapped potential of IoT.
Example 1: Personal computing — from dumb terminals to powerful laptops
In the early days of personal access to computers, people used dumb terminals connected to mini-computers or main frames. All functions available to users were determined by the programs available on those computers. As they were expensive operate, the computing functions available were limited. The computing environments were completely vertically integrated: hardware, operating system, and applications were integrated, and depended on each other.
The advent of personal computers with the PC broke up this vertically integrated environment. While the hardware and the OS were still integrated –the “Wintel” platform-, many manufacturers emerged. In addition, the open application platform enabled a large base of creative programmers and drove a huge expansion of applications available to users. Client / server architecture was an intermediate stage. Ultimately, this development turned the world of computing on its head, with end-users and their lap tops and desk tops driving much of the evolution.
Example 2: Internet access — from walled gardens to the worldwide web
Early Internet services, e.g., America Online, Compuserve, Prodigy, Genie, were vertically integrated “walled gardens”. They controlled the clients and the services available to their users. For some time, they appeared set to completely dominate the online market place. America Online grew so large that they merged with Time Warner, one of the largest content companies at the time.
The advent of HTTP, HTML, and Javascript opened up the vertical integration. By creating an open architecture for content creation, distribution, and interaction, the new vertical segmentation resulted in the open architecture of the World Wide Web. Now, consumers could use any device with any operating system and any browser, to access any content services.
The open content platform unleashed a vast creativity in content providers, generating a tsunami of services, from information services to social media, e-commerce, payments, messaging, and many others. The open browser architecture enabled a large range of devices to access those services. These two developments shifted the initiative driving the evolution of the internet from the service providers to consumers.
Will history repeat itself with the Internet of Things?
For now, most connected consumer devices are connected to walled gardens, e.g., Amazon Alexa, Google Home, and Apple Home Kit, or only to their manufacturers, such as door locks or home surveillance cameras. The majority of industrial IoT devices are connected either to their manufacturers’ clouds, or their owners’ clouds, but generally are not networked among each other, and have little autonomy. They rely on control from the cloud to fulfill their function and create value for their users. In fact, sometimes “edge” devices are considered to be just extensions of the cloud. This is reflected in the term “fog computing” that views edge computers as clouds close to the ground.
The rising power of artificial intelligence, and the rapid increase of their hardware capabilities creates a foundation for edge devices to become more autonomous. At the same time, the cost and time to market for the devices are decreasing rapidly.
The most important development needed for achieving autonomy is seamless interoperability of devices, so that users’ choices are not constrained by the limited ecosystem in which each device operates. At the same time, internetworking of IoT devices to achieve higher level functionality, by integrating collections of devices into the context of their users’ lives, will vastly increase their usefulness and acceptance. Naturally, significant security advances are also needed to achieve breakthrough autonomy.
Industry standards, the equivalent of HTTP, HTML and Javascript on the Worldwide Web, are needed to achieve vertical disintegration required for true device autonomy in the Internet of Things. While it is difficult to predict what standards will finally drive the breakthrough, here are some candidates:
· Edge runtime environment: Docker containers?
· Distribution protocol for containers: some derivative of Kubernetes?
· Radios: software defined radios on top of 5G?
· Standard APIs for sensors and actuators: the actual APIs may be proprietary for some time to come; open intermediating services such as IFTTT can play an important role in accelerating interconnected services.
Increasing autonomy of edge devices and vertical disintegration of IoT services will vastly increase consumers’ choices. Improved inter-networking of devices will increase their usefulness. Breaking the market power of the proprietary platforms will unleash huge new opportunities for independent service providers. It will shift the initiative driving the evolution of the consumer IoT from the proprietary platforms to the consumers. It will unleash the creativity and diversity that device autonomy achieved for the digital information space, leading to an explosion of independent cloud services to provide functions for the newly liberated users.
So how do we make it happen? Would love to have a dialogue. Respond in the comments, drop me an email, or linked in message.