The (IoT) World Is Flat
It was exactly 10 years ago when Thomas Friedman published his bestselling book, the World is Flat: A Brief History of the Twenty-First Century. Ever since Columbus proved that the world is round by reaching the land of “India”, the human race has worked tirelessly to flatten it, according to Friedman. Ten years after the incisive book was published, we continue to witness the flattening transformations of globalization.
Thomas Friedman’s is a flat world of people, brought about by the information revolution. But this flat world of people is soon to create a flat world of things, again powered by the information revolution. The early phase of this flat world of things we are calling “The Internet of Things”.
Connecting things, just like connecting people to the Internet, will turn out to be only the initial stage that will subsequently elicit a much greater transformation to the dynamics of the system, be it social, economical, or technological.
In the flat world of things, computational intelligence will be highly de-centralized, the hierarchies between the “clouds” and the “end points” will blend and disappear, the boundaries between systems will go down and system of systems will emerge to generate much greater automation and efficiency.
The IoT World is (Still) Clumpy
The “flat world theorem” has opponents. For example, Richard Florida, an urban studies theorist counter-argued by saying that “the world is spiky”. He proved it with spiky charts showing that economic activities and innovation are still highly concentrated in selected geographies in the world.
One may also argue that the IoT world is not flat at all — it is clumpy, and is getting even clumpier over time. The world of connected things seems to be a steroid shot served to cloud computing:
- What do I do now with all these data made available from my machines? — send to the cloud.
- How do I make my legacy machine smarter? — connect it to the cloud.
- How to make a number of disparate devices work together as a coherent system? — connect them all to the same cloud.
By 2018, more than three quarters (78 percent) of computing workloads will be processed by cloud data centers; 22 percent will be processed by traditional data centers, according to Cisco. Our favorite technology wonder, the “cloud” is creating more and more concentration of computing power. Where is the flat world?
Recently I heard so much talk of cloud that I told people, when there is a lot of cloud in the sky, one day it is going to rain.
(Forget about the fog thing.)
The centralization de-centralization pendulum
Centralization is a step on the path towards de-centralization. Cloud computing is an early step towards distributed computing.
The human societies, economies, commerce, and institutions have all evolved from distributed, to centralized, and then to de-centralized structures. Book publishing, craft making and manufacturing, journalism, education, you name it.
The 19th and 20th centuries were marked by the extreme centralization of such activities due to the economy of scale brought by the division of labor, made possible by technological advancement, the concentration of capital and infrastructure, and the falling cost of transportation. Such structures are further enforced by the lack of information transparency, which imposes high transaction costs for individuals and smaller entities. This is Globalization 2.0 as pointed out by Friedman.
In Globalization 3.0, the flatter world, your neighbor publishes her movies online; your 15-year-old son is making gadgets and selling on e-Bay. Innovative stuff are getting created everywhere due to the proliferation of hardware incubators; And MOOCs(Massive Open Online Courses) are enrolling more and more students. Next up, let’s see when we witness the decentralization of finance (already started) and farming.
For those familiar with information technology, the equivalent pendulum has swung more than once, from mainframes to PCs and then to cloud. Analogous to what happened in Globalization 2.0 , cloud computing is driven by the economy of scale, the increasing capacity and the falling cost of networking (transport), and the emergence of high performance computing.
But in the flat world of things, you will find computing everywhere. Your stove may learn what is the ideal temperature curve for stir fry; your car may tell that nasty car behind you to maintain safe distance; a group of drilling pumps will work out among themselves how to stabilize collective output. But most importantly, all these are achieved without having to “talk to” a central “Cloud”. This is because the things start to become computers themselves. Not only do they process information internally, they can also communicate with their peers. There is no more master-and-slave — this is a flat world, the democratization for the Internet of Things. Your devices and machines are no more likely to turn to the cloud for regular computation than your 20-year-old son is likely to go to a library looking for answers to his course work.
The Major Flattening Forces
Friedman gave us ten flatteners, out of which three are the foundational forces. Likewise, there are three major “flatteners” for the world of things. Take the first two only and you will get a cloud-dominated IoT world; take all three together and you will leap to the flat IoT world.
First, let me remind you of the three forces underscoring the transition to the flat world of people, in Friedman’s book:
- The collapse of the Berlin Wall and the mass adoption of PCs
- Netscape — the browser that brings the Internet to everyone
- Workflow software — the collections of standards and technologies that allow work to be done wherever it can be done most effectively and efficiently.
So to make a flat world: Step 1, get more information digitized (PCs); Step 2, make information flow between people (Netscape and the Internet for the masses); Step 3, make work flow between people (“workflow software”). Now replace people with things here, and you will get some good hints on how the flat world of things may come about.
Here are the three flattening forces for the IoT world:
- The decoupling of computing from the underlying hardware (virtualization)
- The increasing capacity, prevalence, and speed of networking capabilities
- The ability to disaggregate, distribute and re-combine computing tasks in all environments while maintaining a coherent system
Today, the presence of 1 and 2 without 3 has exactly given us a world of clouds. When computing can be moved around with little constraints from the hardware or the underlying pipes, one of two things may happen — centralization, or decentralization.
Since the tech world has not figured out a way to easily distribute tasks to large numbers of connected things and then re-combine them to create an efficient system, centralization happened instead of decentralization. It may not be a great challenge to arm all energy and power generation equipment with local or embedded computing capabilities. However it is a much bigger challenge to let this distributed network of things work as a collaborative system to balance energy conversion and output. As a result, work gets more and more concentrated because that is the most efficient way known.
Using again the world of people as an analogy, in Globalization 2.0 phase, before the emergence of what Friedman calls “the workflow software”, production and commence became increasingly centralized to gain economy of scale. But after workflow software, your radiology image is taken in the U.S.; the reading is done by a radiologist sitting in India; the medical transcription is performed remotely in the Philippines; and the trouble code from the CT scanner is examined by someone in Germany. They all then come together to produce one outcome in the most cost-effective way. The world flattens after it has become spiky. Value creation went from vertical to horizontal: command and control became connect and collaborate.
The same transition for the IoT world will come when the equivalent of “workflow software” arrives for interconnected things. When work can “flow” across things freely, they will be distributed to wherever and whoever is most efficient. We will find out that the best solution is not always in the cloud, and a group of horizontally connected machines may be able to connect and collaborate. This will drive the 78% of processing from cloud data centers to next to things and inside things.
My optimism says that the flat world of things is coming in the next few years. Aside from the fact that Atomiton is working to provide such technology, I also see the Berlin wall falling for the world of things: that is, the wall between OT and IT will be falling. I see all the other seven flatteners described by Friedman for the world of people happening in the world of things, including outsourcing and supply chaining…
But those should be the topics of my future post…
By Alok Batra,