Distributed technologies are systems of technology composed of many small parts without centralized coordination or control. Key features to distributed systems include the fact that they are typically user-generated, informal networks, involving non-professional producers, with no one really in control of the entire network system. Examples of this would include peer-to-peer file sharing networks on the internet, mesh cellular telephone networks, or swarm robotic systems.1
Our traditional industrial systems of technology are based upon a centralized model design to leverage economies of scale through batch processing, as centrally controlled systems like power generators plants, factories and broadcast media produced technologies and services that were pushed out to end users. Although distributed technologies have always been there on the fringes, today a number of factors are working to fundamentally change this centralized model to a more distributed one, where capabilities and production can also take place at the edges of the network by many different actors.
Factors enabling this include; Firstly, the emergence of alternative technologies like solar cells, wind turbines and 3d printers as increasingly efficient enough to complete and become mainstream products; Secondly, information technology that allows end-users to set up their own networks of coordination and collaboration at very low costs; And lastly, the deregulation and privatization of many previously state-owned monopoly infrastructure industries that is taking place around the world, as previously vertically integrated national systems are being unbundled allowing for a multiplicity of private actors to enter the value chain, as producers, traders, brokers, retailers and many more actors in a complex ecosystem.
A good example of shift from centralized to distributed is mobile telephony. If we look at the emerging infrastructures in rural Africa or Asia today, they often bypass the centralized copper telephone network altogether instead implementing a decentralized cellular network. This helps to demonstrate how distributed technologies thrive particularly in areas with low population that lack the critical mass required for traditional centralized batch processing systems, and also where there is a lack of formal administration and pre-existing incubates. The emergence of renewable smart grids is another example of this. Information systems and distributed technologies are working to fundamentally re-architect the network away from the centrally controlled and operated traditional power grids that delivered electricity to end-users towards a distributed architecture where end-users are central as both consumers and producers, and also as managers of the systems through smart devices and information about prices and consumption.
The same factors that are driving the re-emergence of distributed systems in power grids as viable competitors to the centralized mode are emerging across all domains from digital fabrication in manufacturing to organic farming in agriculture, to car sharing services in transportation and Voice Over IP services within the telephone industry. Distributed systems are no longer the fringe phenomena that they have always been, but increasingly accepted into the mainstream as viable and scalable solutions. These distributed technologies have a number of common features to them, including the fact that they are typically user-generated, informal networks, involve non-professional producers, with no one in control of the entire network system.
Firstly, they are typically user generated. Centrally designed infrastructures are part of a whole paradigm of industrial age organization based around the nation state and bureaucratic, top-down organization, where the end-user is seen as a passive recipient of the service. They are defined as ‘consumers’ and that is essentially their role within the system. These distributed technologies flip this paradigm on its head, actively engaging end-users to become both producers, capable of managing their own resources and capable of setting up their own networks of collaboration through IT.
These distributed networks are informal arrangements. Large mass processing industrial systems like motorways, broadcast media or factories require a significant level of social, political and economic organization over a prolonged period of time through a very formal process of development and management. Due to this, they have a very high threshold to entry as a producer. In contrast to this, these distributed systems are informal with very low barriers to entry. For the cost of a solar panel, individual homes can become power producers. They can also become media producer through new media. They can become manufacturers through digital manufacturing systems. They can become transportation providers through car sharing, all at very low levels of capital investment and organization management, meaning you can do it without needing to be a formal well-defined well-capitalized organization.
There is also a shift from professional to non-professional. The formal centralized model required very high levels of technical capability and specialization in order to produce, operate and maintain. As technologies become commoditized, they become cheaper, more accessible and easier to operate. Distributed manufacturing is an example of this. Digital manufacturing processes such as 3D printers and CNC, numerical controlled technologies were first used by a minority of technicians in factories. As the technology matures and comes to the consumer market, they are increasingly accessible to any end-user without professional expertise in manufacturing or engineering. Again, this goes back to a reduction in barriers to entry. A recent study done by IBM showed that the cost of capital to start a new factory is going to be reduced by 90% in the next decade, which will inevitably drive a democratization of manufacturing. Many distributed technologies like photovoltaic cells are enabled by sophisticated science and engineering but are also accessible for use by the non- technical end-users.
Lastly, no one is really in control of these distributed systems. The centralized infrastructure systems of the nation state were centrally controlled and managed in a hierarchical fashion. The privatization of infrastructure like water, road and rail networks added many more actors to this, as they became managed through market mechanisms. But with true distributed technologies, management of the system may become fully distributed out to the local level of the end-user. In some circumstances, clear ownership may not even apply. Fully distributed technologies are also managed in a distributed self-organizing fashion. Mesh networks and peer-to-peer file sharing are examples of this. Every user supports the provision of the network service. They have a swarm-like dynamic. Control of the system’s behavior and functionality is in the hands of many.
Distributed technologies are also called inverse infrastructure as they have an almost inverse effect on the make-up to our infrastructure to that of the industrial model.2 If you have spent any time in the technology industries, you would have heard the word ‘disruptive’ repeatedly, and it is this structural transformation that is a large part of where this disruption is coming from. Because these alternative technologies simply do not fit into our pre-existing model, and they are not about to simply go away, something has to give. The whole landscape has to change, and that is a scale of disruption that we have not had for a number of centuries; since the Industrial Revolution.
Key to this transformation is designing new frameworks for integrating centralized top-down systems with these bottom-up distributed technologies. All of these previously mentioned factors that differentiate the two architectural paradigms create a tension between a top-down and a bottom-up model. Progressive companies are leading in creating a new architecture that is better suited to this emerging reality, where traditional large centralized players reinvent themselves as platform services providers, providing the infrastructure, protocols and other coordination mechanism for individuals and small organizations to co-create solutions.
Continuing on with the example of the smart power grid, a new model for power grids is emerging called virtual power plants, where centralized wholesale power companies use information technology to integrate and coordinate a number of distributed micro-generators, thus providing a stable supply of energy to the market at the large scale required. This is an example of a hybrid model that may well represent some kind of resolution to the top-down bottom- up core constraint. It is built on the design paradigm of platform technologies that we previously mentioned. Apple’s App Store is a good example of a platform technology. The centralized organization provides the platform for end-users to create the solutions. Other examples might include Zipcar, who provide the platform for car sharing.
These are examples of engineered systems that have managed to integrate a centralized and distributed model to create a highly successful hybrid system. And of course, probably the most successful technology of our time, the Internet, is a classical example of a distributed modular system. With the advent of web 2.0 and technologies like social networking, over a billion people contribute to its making, with highly successful companies like Google working by providing centralized platforms for accessing this mass of distributed user-generated content.
Distributed technologies are having a radical, pervasive and unstoppable effect on our technology landscape. The infrastructure systems of 21st century will not be like that of 20th century, centralized structure, but also it is unlikely to be fully distributed. It will be a much more complex interplay of both. Within the industrial model, we had mainstream centralized systems like factories and motorways that dominated, and we had alternative distributed system that lay on the fringes like organic farming and artisan production. It was a strong dichotomy. The two never really fit together. But with the move of distributed systems into the mainstream, this strong dichotomy is becoming fuzzier. Going forward, there will very likely be a whole spectrum from highly centralized systems like mega cities, to radically distributed systems like Bitcoin, and all kinds of hybrids in-between as we develop new methods for integrating them. As we previously mentioned, this breakdown of the barrier and greater interconnectivity between the two models is a key source of complexity.