7 Reasons Why the Future of Computing is Distributed
The Future of Computing is distributed. The Future of Programming is distributed too. The distributed world is in fact the world of data and to get there data has to be at the center of that world. Since the very invention of the concept of database, over fifty years ago, achieving a data centric world has been a Holy Grail. Finally, as we move to the distributed world we are also finally in a position to arrive at the world with data at the center.
Movies, particularly science fiction movies, show us the future: smart machines, smart sensors, all revolving around data at the center give us incredible knowledge and control over the world around us. Recent developments in chip technologies mean this will all be happening in the very near future — as a result, data will be at the center of our thinking, our decision-making, our society and our company structures. Machine Learning is an important step to put data in the center, however it is only one step. Beyond Machine learning, evolving chip technologies, sensors, and storage require a fundamental change in software architecture, programming, and applications. The good news is that this means we are on the verge of a world with application as servants, resulting in a world of abundance and creativity.
When we put “data” at the center of programming one of the underlying building blocks is Distributed Architecture and Computing. We believe that the edge and the cloud are all a part of a distributed architecture or a hybrid model where the Cloud is just another device in the architecture. More and more we are realizing that the cloud can’t solve all of our problems. Some processing happens on the device and some happens on the cloud. Some decisions happen in the device and some in the cloud. Some storage is located on the device and some on the cloud. With distributed architecture, data will be at the center of the application and the application will be distributed everywhere at the edge. Not processes, Data, thus this will give birth to finally “human” being at the center of everything.
Here are 7 Reasons why the Future of Computing will be Distributed.
1. Building the Data Centered World: Historically the application has been the center of the computing world. Amazon, Facebook and Google put their apps at the center so that they can keep their data all for themselves, carefully controlling and doling out what you can see. Once Machine Learning lets computers see and hear, it becomes possible to build even more sophisticated world models. The information in those models cries out to be shared among all applications wherever they are. The answer is to invert our application architecture so that data moves to the center. A data centered world is an intrinsically distributed world. Apps all over the world, running on a wide variety of machines all contribute to and share the data and world models now available to us all. In the new world if I build a social graph, everybody gets to access it directly. Instead of being a carefully guarded proprietary asset, that social graph becomes a shared engine for change. The shift to data centered apps will start inside companies and inside smart machines. In a factory, instead of separate silos for purchasing, design, scheduling, manufacturing, inventory, shipping and so on, the system shifts to a single shared database that always has a complete picture of the entire process. In a car, data from a wide variety of sensors, from both inside the car and the outside world are all collected in a single world model that is used by action engines located all through that same car.
2. Sensors, sensors everywhere: With chips getting small and power efficient, we will have sensors everywhere. Not only in our smartphones, in our cars or on our wrists, but also in our shoes, in our clothes, everywhere in our surroundings, working spaces like factories and offices and houses. Data collected with these sensors will be processed, allowing real-time decisions to be given right on the device. Some of data needs to be “curated” and synched to other devices and replicated for coherent data and processing. The extension of Distributed Architecture allows truly distributed, peer-to-peer communication that can be as real-time as needed.
3. Unbundling and Deconstruction of the Computer — The Network Is The Computer: Bundling and unbundling are two fundamental forces for business and societal change. Smartphones bundled together phones, computers and cameras creating a huge new market. Originally mainframes bundled together computing, storage and even networking. The distributed world is all about the unbundling of what we used to know as the computer.
Until today we worked with a single computer that contained our storage, our apps, and our entire experience. Even with the internet, the smartphone may be a computer, but really it is a terminal accessing the bundled computer service in the cloud. Distributed computing is changing that in a very large way. First each of us is increasingly at the center of a personal cloud of separated but cooperating computers: our phone, our watch, the air pods in our ears, the drone. When seven or eight cars negotiate how to handle an intersection, where is the computer — it is the computers in all the cars working together.
Computing is everywhere. Storage is everywhere too. The robot surgeon has its own huge database as does the car, the delivery drone and the autochef. It is the fact that data has moved to the center of system that allows all the pieces to work together. Tomorrow’s architectures that will choose the most efficient place to put the Computing and the best place for the storage — making it possible for real time data and real time decision making. In this distributed world of the future, there is no bundled computer any more: the network, with its billions of specialized computers is the computer.
4. Next Billion Entrepreurs who will write applications without the Goal to be the Next Facebook: Worldwide, IT departments are becoming smaller and smaller. They are at a stage that they can’t get smaller any more. Yet, organizations all need to go through a major transformation putting “Data” in the center, instead of the current process oriented application structure. Whether you call this change the AI revolution or not, this change is happening. For example, factories are going through this massive change, embedding machine-learning mechanisms and ”data” based real time decision-making. Who will write the necessary applications needed to make this change? How? Using which tools?
Let’s take developers as another example. There are only 15 big-scaled networks in the world built by a small army of centralized developers. However, there are 10,000,000 application opportunities out there, which are still useful. Not every application will have the biggest network effects such as Facebook, Twitter or Airbnb. We are in need of applications that serve a small need and will not be used by billions of people and we need the developers to build them. We need an environment with a new generation of applications that are:
a) bootstrapped/sponsored in a way,
b) Written/ programmed easily using the new architecture,
c) Data is in the center of the architecture,
The need for these millions of applications is another reason why the world will be distributed: it is just not possible for all these applications to be written centrally, particularly since the people at the center simply will not understand the needs of the machines and users at the edge.
5. Trust in the Centers? Developers are tired of depending on big networks and depending on centralized systems. They are tired of paying 30% to the centralized platforms. They are tired of the threat of APIs being unavailable, the threat of being shut off or limiting their growth. As a starting point, in a distributed application world every developer can build exactly the app most appropriate for the need. Since the applications run on distributed machines, there are no fees to be paid, and the choice of api’s is all that of the developer. The only challenge is how to come up with the marketing and distribution platforms that help developers reach their customers in fast and fluid way. Again, the distributed platform has a chance to come to the rescue in the form of communities and community platforms.
6. Industries become Information Economies: It’s not just computers: entire industries are being unbundled — in the process they are becoming literally information economies in their own right. When industries become Information Economies, meaning software businesses, massive changes happen: with more innovation, costs go way down and democratization arrives. Transportation and Energy are two good examples of this transformational process. As the transformation occurs, in one industry after another, these trillion dollar industries are all turning into distributed, software driven networks.
Historically energy was one of the most centralized businesses with massive generation facilities supplying power to entire states and cities. Today, solar panels and wind technologies allow everyone to generate their own electricity and store it in their own batteries. The challenge for individuals is how to deal with the peaks and valleys. The answer is a distributed architecture with real time decision-making and device-to-device communication so that power can be bought, sold, and distributed to make it work for everyone automatically.
The transportation industry is going through an even more massive shift to Information Economy than Energy. Of course we all know that cars and even ebikes today are built around internal computers and networks. Until now, though, the basic nature of the industry — huge companies selling cars to millions of individual owners — has not changed since the beginning. Today though, even that ownership model may disappear. Services like Lyft and Uber allow us to think about calling for cars whenever we need them. When the cars become self-driving, even the driver will disappear from the equation. In this new world, users have far more flexibility and choice than ever before. Instead of being stuck with the single size and type of car you bought and owned, now you can use an eBike, call for a Lyft, or arrange for a small truck that you pick up and leave wherever it best suits you. At that point, the entire system is one huge distributed network with computers everywhere all working together. In this new world, data is at the center, and information flow is as ubiquitous and central as the movement of physical machines.
7. More Disintegration — A Distributed Incentive and Transaction Layer: The existing payment and transaction system is one of the most centralized parts of the modern world. That same payment system, with banks at the center, is often a bottleneck at multiple levels. By embedding the economic model in the Internet we can bring a new business model and incentives to developers solving many bootstrapping problems. Blockchain systems that enable payment directly from the protocol level — including the potential for micropayments — is liberating to developer networks. Ethereum as a computational resource, Filecoin as a global database resource are early examples to how incentives and applications can be unbundled from current payment models and bundled to new contract models. Beyond payment, cryptocurrency allows us also to try new contract, governance models and financial models, resulting in the disruption of many industries helping drive them to Information Economies.
The transformation of industries into information economies both requires and drives a shift in the underlying ownership and economic model too. In transportation for example, think about the shift from ownership to hailing to subscription and eventually to a contract based model. The shift to hailing is called Servitization, from a product to a service. Subscriptions are about Securitization and the shift from service to capital asset. Finally, when we get to contracts, we have Tokenization: capital assets become tokens. This is happening today and we will see more of it in the very near future in personal transportation from bikes to cars as well as public and logistics.
We can extend this train of thought to many information industries. Many industries once become an Information economy can be governing the model in the future once we bridge offline value offering with online coin, which will allow many distributed applications to be written for these industries. Examples are endless from finance applications to energy, transportation and healthcare applications.
As we conclude, let’s think about Bitcoin from an energy perspective. In a way, Bitcoin is a coin to incentivize electricity consumption. That is another reason why a country like China, which provides free electricity, became a big force with server farms providing free electricity for Bitcoin mining.
How about a “Green Coin” which will incentivize solar energy production and sharing?
What else will be disintegrated and distributed? The answers are no longer million dollar questions but in many cases trillion dollar questions, giving birth to new ecosystems.