A trending topic in technology is the debate between centralization & decentralization. In Decentralization in all the things, Mike Masnick aims to find a middle ground by arguing for an integrated approach that can be summed up as Centralized Infrastructure with Decentralized Innovation.
The argument is this: Centralization should be used to establish core infrastructure, ensuring uniformity, efficiency, and scale. This infrastructure should then serve as a foundation upon which highly responsive decentralized entities can innovate. For example, the internet itself is built on standardized protocols (a form of centralized infrastructure) that allow a myriad of decentralized applications and services to exist.
While I agree with the core premise, I don’t think that his definitions of centralization or decentralization are accurate or useful. Masnick cites US road networks and internet protocols as examples of ‘centralized’ infrastructure. Later in the article, he mentions Facebook as an additional example of a centralized entity.
The definition of centralization which he is seemingly using here is somewhat similar to ‘monopoly’ in that there is only one of that thing, or a single standard. Road or rail networks are often-cited examples of ‘natural monopolies:’
*A natural monopoly is a distinct type of monopoly that may arise when there are extremely high fixed costs of distribution, such as exist when large-scale infrastructure is required to ensure supply. Economics Online
However, both of these definitions miss the fact that there is an enormous difference between a public good, which is owned and/or managed collectively, vs a private company or good, which is often owned and governed by a small number of people exclusively for the purpose of increasing profits.
Road and rail systems are usually public goods, owned and managed by governments for the benefit of society. Internet infrastructure such as the Internet Protocol, Simple Mail Transfer Protocol (SMTP) & Ethereum Protocol are also public goods. These digital protocols, although ‘centralized’ by Masnick’s definition, are largely considered decentralized because they are open infrastructures that anyone can use.
The Ethereum Protocol and blockchain protocols more generally also aim to decentralize power by eliminating the need for a single source of truth and developing distributed governance models. To varying degrees, democratic governments could also be said to decentralize power, as they at least theoretically have mandates to listen to and serve their people. This is not the case with market monopolies like Facebook.
In terms of internet infrastructure, Masnick cites two examples of decoupling fibre-optic networks from internet service providers: a publicly-owned network built by Ammon, Idaho, and the breaking up of BT by the UK government.
I don’t know much about either case, so I might be pick-and-choosing data points here, but from my research, it seems like there is a big difference between the outcomes of the two case studies. The Ammon model was implemented to serve the public interest, and appears to have lowered the cost of internet, increased speeds, and brought economic prosperity to the area. The model has been replicated in other municipalities, including Idaho Falls. Source
On the other hand, BT spun off its infrastructure division, Openreach, into a separate company following a regulatory review in 2005. Even though Openreach is independent, is is still wholly owned by the BT group and generates 35% of their profits. The company has been accused of abusing its monopoly to reduce infrastructure investments and service quality. Source
It’s clear that we do need standardized protocols and infrastructure to drive innovation, as mentioned in 3. Standardization of Components. However, the goals and governance structures of these institutions matter. The real rift is not between ‘centralized’ or ‘decentralized’ systems but between closed and open systems, between core structural systems managed by the few for the benefit of the few or systems run for our collective benefit.
As the internet shows, if we don’t keep core infrastructure open and interoperable, walled gardens will arise that create worse outcomes for everyone. The promise of decentralization is ironically exactly what Masnick terms ‘centralization’: the ability to build open, interoperable infrastructure that drives innovation.
Dive Deeper
Topic relates to:
Blockchains Protocols Institutions
Further reading:
Decentralization in all the things Guest Talk with Simone Cicero | Perspectives on the Platforms vs Protocols Debate: A Reality Check The Meaning of Decentralization